From: dcurry@silo.csci.unt.edu (David Mason Curry) Newsgroups: alt.folklore.computers Subject: Is VMS taboo, or what? Date: 24 Nov 1998 16:02:08 GMT Organization: University of North Texas Lines: 10 Message-ID: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: silo.csci.unt.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!news.maxwell.syr.edu!cs.utexas.edu!news.unt.edu!silo.csci.unt.edu!dcurry I am too young to have seen glorious things such as Multics, ITS, or TOPS-20, but I love reading about them. However, one thing that I have noticed: *nobody* even mentions VMS! Why? I worked with VMS for years in college and then when I went to work for TI (HARM missile program) and I like VMS far more than any un*x system I've had the displeasure of using. So tell me: Is VMS a toy compared to other operating systems? Cheers, David Curry ###### From: "Jack Peacock" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 24 Nov 1998 10:19:09 -0800 Organization: Simco Lines: 30 Message-ID: <73etgc$nrj$1@supernews.com> References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: 207.168.124.145 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Trace: 911931724 TWNY2XU9G7C91CFA8C usenet57.supernews.com X-Complaints-To: newsabuse@supernews.com X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail David Mason Curry wrote in message <73ela0$51h@hermes.acs.unt.edu>... >I am too young to have seen glorious things such as Multics, ITS, or >TOPS-20, but I love reading about them. However, one thing that I have >noticed: *nobody* even mentions VMS! Why? I worked with VMS for years >in college and then when I went to work for TI (HARM missile program) >and I like VMS far more than any un*x system I've had the displeasure >of using. So tell me: Is VMS a toy compared to other operating systems? > You have it backwards. Other OSes are toys compared to VMS. The reason you don't see much talk about it outside the VMS newsgroups is that there aren't very many problems with VMS, not like DOS, NT, Win95, etc. VMS installations tend to be professionally managed sites rather than hobby machines. I still do all my development on a mixed VAX and Alpha VMS cluster. It's always up, never lost files, good manuals with just about all the information you need, and you can still get printed documentation. The trend now is to merge VMS into NT networks, which is the path we have chosen at work, using a network of Alpha NT and VMS servers, plus a few VAX VMS servers to maintain old legacy programs. Jack Peacock ###### From: Victor Eijkhout Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 24 Nov 1998 16:26:18 -0500 Organization: University of Tennessee Lines: 14 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: prancer.cs.utk.edu X-Newsreader: Gnus v5.5/Emacs 20.2 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newspeer1.nac.net!news.maxwell.syr.edu!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!cam-news-hub1.bbnplanet.com!cpk-news-feed4.bbnplanet.com!cpk-news-feed1.bbnplanet.com!news.gtei.net!utk.edu!not-for-mail rivie@rivie.daautah.com (Roger Ivie) writes: > >noticed: *nobody* even mentions VMS! Why? I worked with VMS for years > There's a very active VMS newsgroup: comp.os.vms. Translation: go away. :-) -- Victor Eijkhout "Too many textbooks and discussions leave children free to make up their minds about things." [Fundie Mel Gabler] ###### From: rivie@rivie.daautah.com (Roger Ivie) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 24 Nov 1998 16:31:51 GMT Organization: XMission http://www.xmission.com/ Lines: 17 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> Reply-To: rivie@daa-utah.com NNTP-Posting-Host: logan13.modem.xmission.com X-Trace: news.xmission.com 911925111 18463 166.70.3.205 (24 Nov 1998 16:31:51 GMT) X-Complaints-To: abuse@xmission.com NNTP-Posting-Date: 24 Nov 1998 16:31:51 GMT X-Newsreader: slrn (0.8.8.2 UNIX) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!xmission!nnrp.xmission!rivie In article <73ela0$51h@hermes.acs.unt.edu>, David Mason Curry wrote: >I am too young to have seen glorious things such as Multics, ITS, or >TOPS-20, but I love reading about them. However, one thing that I have >noticed: *nobody* even mentions VMS! Why? I worked with VMS for years >in college and then when I went to work for TI (HARM missile program) >and I like VMS far more than any un*x system I've had the displeasure >of using. So tell me: Is VMS a toy compared to other operating systems? There's a very active VMS newsgroup: comp.os.vms. -- Roger Ivie Design Analysis Associates 75 West 100 South Logan, UT 84321 mailto:rivie@daa-utah.com phoneto:(435)753-2212 ###### From: dpeschel@u.washington.edu (D. Peschel) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 24 Nov 1998 16:58:55 GMT Organization: University of Washington, Seattle Lines: 39 Message-ID: <73eokf$16lk$1@nntp1.u.washington.edu> References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: saul9.u.washington.edu X-Trace: nntp1.u.washington.edu 911926735 39604 (None) 140.142.17.39 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: dpeschel Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!192.220.250.21!netnews1.nw.verio.net!netnews.nwnet.net!news.u.washington.edu!dpeschel In article <73ela0$51h@hermes.acs.unt.edu>, David Mason Curry wrote: >of using. So tell me: Is VMS a toy compared to other operating systems? No, it isn't. I don't think many people would say it is. It may be idio- syncratic and annoying, but it's not a toy. Some people on this group use VMS for work _and_ play. But I've only used it for work (like you seem to have). And I've never had decent privileges, or seen the source, or really done a lot of hacking. As I said above, at my "peon" level I think it can really be a @$!@$# pain to use. As you can see, all those factors in themselves would be enough to make me a UNIX bigot (which I am). But when you also factor in VMS' proprietary nature, its "second-class" status regarding the UNIX software that everyone wants, and its expense (even now), not to mention the fact that DEC doesn't even really _exist_ any more, it's just more fruitful to talk about UNIX. This is making a complicated issue simple... I don't know which of those factors above influenced the others. Certainly, there are some fun things to do with VMS (including digging out old software, e.g., from DECUS). But you need an Alpha (or possibly a VAX) to do them! Not to mention the proper DEC graphics terminals, etc. Also, finding that software is (to me) much harder than with UNIX, because VMS never completely embraced the Internet, I think. I'd also say that's the reason why you never hear about or can use TOPS anymore. ITS and SAIL and WAITS are even less accessible because they were local. Many of the Multics people hang out here, and there's a fair amount of information (but no software/hardware) freely available about Multics now. VMS' ancestors (e.g., RSX, RT-11) are actually _more_ fun to play with because they are more manageable in size and complexity. Some other people here may agree/disagree with me. Allison? Multics people? Rich Alderson? Anyone? Anyone? :) And I'm sure you will get a different story on comp.os.vms. -- Derek ###### Message-ID: <365B18D7.20A5AC1@zen.co.uk> From: "Alan J. Wylie" Organization: X-Mailer: Mozilla 4.5b2 [en] (WinNT; I) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 7 Date: Tue, 24 Nov 1998 20:36:39 +0000 NNTP-Posting-Host: 212.23.8.2 X-Trace: news-reader.bt.net 911939801 212.23.8.2 (Tue, 24 Nov 1998 20:36:41 BST) NNTP-Posting-Date: Tue, 24 Nov 1998 20:36:41 BST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!diablo.theplanet.net!btnet-peer!btnet-feed1!btnet!news-reader.bt.net!not-for-mail "D. Peschel" wrote: [SNIP] > because VMS never > completely embraced the Internet, I think. Not just failed to embrace, but invented their own broken alternative to TCP/IP - DECNET ###### From: Brian Macke Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 24 Nov 1998 21:09:07 GMT Organization: Strangelove Networking Lines: 19 Message-ID: <73f79j$rit$1@sword.avalon.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> Reply-To: macke@strangelove.net NNTP-Posting-Host: ripper.strangelove.net X-Newsreader: NN version 6.5.0 (NOV) Originator: macke@ripper.strangelove.net Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!204.71.106.34!avalon.net!ripper.strangelove.net!macke "Alan J. Wylie" writes: >"D. Peschel" wrote: >[SNIP] >> because VMS never >> completely embraced the Internet, I think. >Not just failed to embrace, but invented their own >broken alternative to TCP/IP - DECNET ...don't forget the.. frustration associated with attempting to get a VMS box onto any IP network, not just the Internet. Did you have Wollowong, UCX, Multinet...? What were the other 'flavours' of TCP/IP you could get for VMS? I know there were at least five. -- -Brian James Macke macke@strangelove.net The Prophet "In order to get that which you wish for, you must first get that which builds it." -- Unknown ###### From: gerglery@usa.net (Fluffy) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 24 Nov 1998 21:23:57 GMT Organization: Ruler of Usenet (all) Message-ID: <73f85d$bfp$1@meowhost.meow.invalid> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> X-Trace: mail2news.demon.co.uk 911944464 mail2news:27489 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!out2.ibm.net!meowhost.meow.invalid X-P-Meow: Meow Mail-Copies-To: never X-URL: http://members.tripod.com/~gerglery/ X-Newsreader: slrn (0.9.5.3-canlock UNIX) Lines: 12 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Brian Macke wrote: > ...don't forget the.. frustration associated with attempting to get a VMS > box onto any IP network, not just the Internet. Did you have Wollowong, Wollongong. > UCX, Multinet...? What were the other 'flavours' of TCP/IP you could get > for VMS? I know there were at least five. Tek/CMU (that's the free one). TCPware (Process now owns Multinet too, after Cisco bought TGV.). ###### From: "Jack Peacock" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 24 Nov 1998 20:27:58 -0800 Organization: personal use Lines: 31 Message-ID: <73g0uo$o0k$1@supernews.com> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> NNTP-Posting-Host: 207.168.124.157 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Trace: 911968024 TWNY2XU9G7C9DCFA8C usenet76.supernews.com X-Complaints-To: newsabuse@supernews.com X-Newsreader: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.wirehub.nl!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail Brian Macke wrote in message <73f79j$rit$1@sword.avalon.net>... >"Alan J. Wylie" writes: > >>"D. Peschel" wrote: >>[SNIP] >>> because VMS never >>> completely embraced the Internet, I think. > >>Not just failed to embrace, but invented their own >>broken alternative to TCP/IP - DECNET > >...don't forget the.. frustration associated with attempting to get a VMS >box onto any IP network, not just the Internet. Did you have Wollowong, >UCX, Multinet...? What were the other 'flavours' of TCP/IP you could get >for VMS? I know there were at least five. >-- First, let's not forget DECnet goes back quite a way, mid 70's, before Ethernet even got started. Back then TCP/IP was an obscure academic protocol. DECnet worked reliably, connecting most of DEC's hardware. A 1Mbit connection on PDP-11's was not shabby at all for the times. I put a VMS cluster on TCP/IP using UCX with no problems, works fine. I've done it at several customer sites too, mostly connecting to Unix machines via NFS. Works fine doing automated file transfers to VMS. Jack Peacock ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 25 Nov 1998 05:59:47 GMT Organization: Arargh!! Lines: 21 Message-ID: <365c9b1e.101247987@news.mcs.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!cpk-news-hub1.bbnplanet.com!chicago-news-feed1.bbnplanet.com!news.gtei.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On 24 Nov 1998 16:58:55 GMT, dpeschel@u.washington.edu (D. Peschel) wrote: >In article <73ela0$51h@hermes.acs.unt.edu>, >David Mason Curry wrote: > >>of using. So tell me: Is VMS a toy compared to other operating systems? > >No, it isn't. I don't think many people would say it is. It may be idio- >syncratic and annoying, but it's not a toy. > >Some people on this group use VMS for work _and_ play. But I've only used it >for work (like you seem to have). And I've never had decent privileges, or >seen the source, or really done a lot of hacking. As I said above, at my >"peon" level I think it can really be a @$!@$# pain to use. I would say that you are being polite. ^^^^^^^^^ arargh >-- Derek ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 25 Nov 1998 05:59:48 GMT Organization: Arargh!! Lines: 14 Message-ID: <365d9b9c.101373846@news.mcs.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!cpk-news-hub1.bbnplanet.com!cam-news-hub1.bbnplanet.com!chicago-news-feed1.bbnplanet.com!news.gtei.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On Tue, 24 Nov 1998 20:36:39 +0000, "Alan J. Wylie" wrote: >"D. Peschel" wrote: >[SNIP] >> because VMS never >> completely embraced the Internet, I think. > >Not just failed to embrace, but invented their own >broken alternative to TCP/IP - DECNET ah yes, the (*&% that changes the MAC address of the card rather than use whats there arargh ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: "Jack Peacock"'s message of Tue, 24 Nov 1998 20:27:58 -0800 Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> Date: Wed, 25 Nov 1998 20:32:50 GMT Lines: 29 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!netnews.com!ix.netcom.com!netcom!alderson In article <73g0uo$o0k$1@supernews.com> "Jack Peacock" writes: >First, let's not forget DECnet goes back quite a way, mid 70's, before >Ethernet even got started. Ethernet dates from around 1971 or 1972, as does the first version of DECNET. There were *lots* of research projects thinking about how to connect computers back then. DECNET was just a commercially supported one--that only worked on one vendor's computers. (You know, the sort of thing we *criticize* IBM for doing? Bisync. SNA.) But Ethernet vs. DECNET is irrelevant; it's not the media (DECNET was made to run on Ethernet fairly early), it's the protocol suites. >Back then TCP/IP was an obscure academic protocol. Back then, TCP/IP didn't exist. There was NCP (for 56Kbit phone lines--you know, the ARPANET?) for long-distance networking, and there were things like ChaosNET (MIT) and PUP (Xerox PARC) for internetworking local-area networks-- but *all* of these were research efforts, not commercial propositions. DECNET locked you into DEC equipment; TCP/IP, when it came along, did not. *That* is the issue between them, *not* whether a DEC-only shop could get along without TCP/IP. -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: Kin Hoong CHUNG Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 27 Nov 1998 09:55:00 GMT Organization: University of New South Wales Lines: 13 Message-ID: <73lstk$ku1$1@mirv.unsw.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73f85d$bfp$1@meowhost.meow.invalid> NNTP-Posting-Host: alpha.maths.unsw.edu.au User-Agent: tin/pre-1.4-980618 (UNIX) (OSF1/V4.0 (alpha)) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!sunqbc.risq.qc.ca!news.uow.edu.au!news.usyd.edu.au!unsw.edu.au!not-for-mail Fluffy wrote: : Brian Macke wrote: :> ...don't forget the.. frustration associated with attempting to get a VMS :> box onto any IP network, not just the Internet. Did you have Wollowong, : Wollongong. Is that the reason for the wombat reference? Why that Australian city anyway? Cheers, Kin Hoong ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 27 Nov 1998 22:12:11 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 9 Message-ID: <3662237d.19279614@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On 24 Nov 1998 16:02:08 GMT, dcurry@silo.csci.unt.edu (David Mason Curry) wrote: > So tell me: Is VMS a toy compared to other operating systems? No. Toys are things that are fun to play with. Nothing could be farther from the truth regarding VMS. It sucks so hard that it actually hurts. ###### From: Chris Pitzel Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 27 Nov 1998 23:26:40 -0600 Organization: Engineering, University of Saskatchewan Lines: 36 Message-ID: <365F8990.12A351F8@nospam.usask.ca> References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: janus2-2.usask.ca Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.5 [en] (WinNT; U) X-Accept-Language: en,fr-CA,fr Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!dca1-hub1.news.digex.net!digex!news-feed.fnsi.net!enews.sgi.com!decwrl!tribune.usask.ca!not-for-mail > noticed: *nobody* even mentions VMS! Why? I worked with VMS for years VMS, while still being far superior to any other operating systems in terms of stability, uptime, and security, was never marketed properly. Ultimately, the poor management of VMS sales and marketing led to DEC's takeover by an even-more despised company, Compaq. Most DEC-buying people I have talked vow never to buy another DEC/Compaq product due to the takeover.. If some Alphaserver or VAX needs replacing, they will do everything possible to not end up in the situation of having to buy a "Compaq Alphaserver" (I noticed the other day that Compaq's actually starting to advertise them in banner ads and in magazines..). If DEC had marketed VMS properly, we'd all be sitting infront of VMS running machines today (or at least a good number of us.). Over time, they would've reworked the GUI into something nice and pretty (like WindowsNT is today..), added a working and half-decent POSIX subsystem, and added in all of the multimedia components and networking prowess and ease of use that is expected of an OS today. > in college and then when I went to work for TI (HARM missile program) > and I like VMS far more than any un*x system I've had the displeasure Yes, I like VMS very much still. They don't give me account on the VMS machine around here anymore, yet I used to use it for years. > of using. So tell me: Is VMS a toy compared to other operating systems? Not really. It's one of the most robust operating systems out there. Nobody really likes talking about it, though, because that would imply talking about (now-disgraced) Digital and (despised by DEC fans) Compaq.. The Compaq takeover was the worst possible outcome for Digital, but Digital unfortunately lacked any marketing talent and couldn't even make enough money to survive out there in the real world.. ###### From: huge@nospam.demon.co.uk (Hugh Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 28 Nov 1998 11:02:52 GMT Organization: Piglet's Pickles and Preserves Message-ID: <73ol8s$ms7@axalotl.demon.co.uk> References: <365F8990.12A351F8@nospam.usask.ca> Reply-To: huge@nospam.demon.co.uk NNTP-Posting-Host: axalotl.demon.co.uk X-NNTP-Posting-Host: axalotl.demon.co.uk:158.152.24.143 X-Trace: news.demon.co.uk 912254648 nnrp-08:22132 NO-IDENT axalotl.demon.co.uk:158.152.24.143 X-Complaints-To: abuse@demon.net x-no-archive: yes Lines: 16 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!axalotl.demon.co.uk!axalotl!usenet In article <365F8990.12A351F8@nospam.usask.ca>, Chris Pitzel writes: >Over time, >they would've reworked the GUI into something nice and pretty (like >WindowsNT is today..), And there I was thinking you were serious, until you got to this bit .... :o) -- "The road to Paradise is through Intercourse." The uk.transport FAQ; http://www.axalotl.demon.co.uk/transport/FAQ.html [Substitute "axalotl" for "nospam" to email me] ###### From: tangentSPAMCATCHER@cyberport.com (Warren Young) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 28 Nov 1998 17:41:09 GMT Organization: none Lines: 24 Message-ID: <366034e9.2284371302@news.cyberport.com> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73f85d$bfp$1@meowhost.meow.invalid> <73lstk$ku1$1@mirv.unsw.edu.au> NNTP-Posting-Host: 56k26-118.cyberport.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: macaw.cyberport.com 912274875 16464 204.134.118.26 (28 Nov 1998 17:41:15 GMT) X-Complaints-To: abuse@cyberport.com NNTP-Posting-Date: 28 Nov 1998 17:41:15 GMT X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-dc.gip.net!news.gsl.net!gip.net!news.winstar.net!winstar!news.good.net!set!news.cyberport.com!not-for-mail Kin Hoong CHUNG wrote: >Fluffy wrote: >: Brian Macke wrote: >:> ...don't forget the.. frustration associated with attempting to get a VMS >:> box onto any IP network, not just the Internet. Did you have Wollowong, > >: Wollongong. > >Is that the reason for the wombat reference? Why that Australian city >anyway? Because a group of people at the University of Wollongong wrote a TCP/IP stack, and then went commercial with it. I'm sketchy on the details, but Peter Salus' fine book _A Quarter Century of UNIX_ touches on this. Up to a few years ago, they even had a PC stack, but then Microsoft killed off the third-party Winsock stack market, and Wollongong was swallowed by some bigger company, whose name I cannot recall. = Warren -- http://www.cyberport.com/~tangent/ = ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m -- Don't judge a book by its movie. ###### From: gerglery@usa.net (Fluffy) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 29 Nov 1998 02:03:10 GMT Organization: Ruler of Usenet (all) Message-ID: <73qa0u$uo3$1@meowhost.meow.invalid> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73f85d$bfp$1@meowhost.meow.invalid> <73lstk$ku1$1@mirv.unsw.edu.au> <366034e9.2284371302@news.cyberport.com> X-Trace: mail2news.demon.co.uk 912310016 mail2news:3467 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!out2.ibm.net!meowhost.meow.invalid X-P-Meow: Meow Mail-Copies-To: never X-URL: http://members.tripod.com/~gerglery/ X-Newsreader: slrn (0.9.5.3-canlock UNIX) Lines: 6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!woodstock.news.demon.net!demon!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Warren Young wrote: > Wollongong was swallowed by some bigger company, whose name I cannot > recall. They ended up as part of Attachmate, which still sells Pathway etc. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 29 Nov 98 12:21:16 GMT Organization: UltraNet Communications, Inc. Lines: 26 Message-ID: <73rfg5$dku$4@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> NNTP-Posting-Host: d14.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 29 Nov 1998 12:42:45 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news.ultranet.com!d14 In article , alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >In article <73g0uo$o0k$1@supernews.com> "Jack Peacock" >writes: > >>First, let's not forget DECnet goes back quite a way, mid 70's, before >>Ethernet even got started. > >Ethernet dates from around 1971 or 1972, as does the first >version of DECNET. I don't remember it being that early. What I do remember is that those DECNET folks kept dithering on about the specs for so long that the TOPS10 folks got impatient and wrote their own--it's called ANF-10. DECNET didn't implement all the stuff that was in ANF-10 until (I think) Phase IV. The ANF-10 developers would go to 198x DECNET review meetings, mention in an offhand way, "Oh, yea, we did that in 1975", to the great disbelief of the VMSers. There...got that off my chest :-). /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 29 Nov 98 12:24:04 GMT Organization: UltraNet Communications, Inc. Lines: 19 Message-ID: <73rflc$dku$5@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> NNTP-Posting-Host: d14.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 29 Nov 1998 12:45:32 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news.ultranet.com!d14 In article <3662237d.19279614@news.home.ibert.com>, martin@ibert.com (Martin Ibert) wrote: >On 24 Nov 1998 16:02:08 GMT, dcurry@silo.csci.unt.edu (David Mason >Curry) wrote: > >> So tell me: Is VMS a toy compared to other operating systems? > >No. Toys are things that are fun to play with. > >Nothing could be farther from the truth regarding VMS. It sucks so >hard that it actually hurts. Now, now. One should be kind to the senile [smiling emoticon here showing its great bias]. /BAH Subtract a hundred and four for e-mail. ###### From: Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 29 Nov 1998 15:21:14 -0500 Organization: Kettering University (formerly GMI E&MI) - Flint MI Lines: 45 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> NNTP-Posting-Host: nova.kettering.edu Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <73rflc$dku$5@strato.ultra.net> Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!24.130.1.14!lsnws01.we.mediaone.net!24.131.1.12!denws01.mw.mediaone.net!news.gmi.edu!nova.kettering.edu!lee1089 On Sun, 29 Nov 1998 jmfbahciv@aol.com wrote: > In article <3662237d.19279614@news.home.ibert.com>, > martin@ibert.com (Martin Ibert) wrote: > >On 24 Nov 1998 16:02:08 GMT, dcurry@silo.csci.unt.edu (David Mason > >Curry) wrote: > > > >> So tell me: Is VMS a toy compared to other operating systems? > > > >No. Toys are things that are fun to play with. > > > >Nothing could be farther from the truth regarding VMS. It sucks so > >hard that it actually hurts. > > Now, now. One should be kind to the senile [smiling emoticon here > showing its great bias]. From what experience I've had (three MicroVAXen servers) VMS is a great operating system. It's very stable[1], relatively non-bloated and an overall decent system. However, it's command syntax is very, very bloated and complex as compared to Unix. For example paths are disk:[dir1.dir2.dir3]file.ext;n where n is a version number. Also the commands are lengthy (DIRECTORY for example) although they can be abbreviated until they aren't ambiguous. For example DIRECTORY can be abbreviated to DIR but not to D. Also the switches are preceded by slashes such as DIR/NEW. Combining the lot you can easily get multiline monstrosities. [1] I once saw a VMS server run completely out of disk space. Not one single block was left. However, the server continued to run (albeit somewhat slowly) and once some space was freed on the hard drive it ran normally again. I must say that I was impressed but not enough to overcome my dislike of the command syntax. ____________________________________________________________________________ | "A little nonsense now and then, | "If it walks out of the fridge, let Is relished by the wisest men." | it go" -- John Dougherty --W.W. | "If it loves you it will come back." | -- Ian Davis __________________________________|_________________________________________ Theta Xi Kappa Sigma ###### From: Joseph Allen Dane Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 30 Nov 1998 08:15:53 -1000 Organization: University of Hawaii Lines: 25 Sender: jdane@honlab.nmfs.hawaii.edu Message-ID: <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> NNTP-Posting-Host: honlab.nmfs.hawaii.edu Mime-Version: 1.0 (generated by tm-edit 7.106) Content-Type: text/plain; charset=US-ASCII X-Newsreader: Gnus v5.6.9/Emacs 19.33 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!nyd.news.ans.net!newsjunkie.ans.net!newsfeeds.ans.net!news.lava.net!news.Hawaii.Edu!news jmfbahciv@aol.com writes: > In article , > alderson@netcom.netcom.com (Richard M. Alderson III) wrote: > >In article <73g0uo$o0k$1@supernews.com> "Jack Peacock" > > >writes: > > > >>First, let's not forget DECnet goes back quite a way, mid 70's, before > >>Ethernet even got started. > > > >Ethernet dates from around 1971 or 1972, as does the first > >version of DECNET. > > I don't remember it being that early. What I do remember is ... There was a paper on Ethernet published by the PARC people in CACM in July, 1976. This was probably not the first time Ethernet was ever mentioned, but it dates it somewhere around that time, I'd say. -- joe ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 30 Nov 1998 23:40:28 +0100 Organization: My own Private Self Lines: 33 Sender: neil@chonsp.franklin.ch Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> X-Newsreader: Gnus v5.3/Emacs 19.34 Joseph Allen Dane writes: > > jmfbahciv@aol.com writes: > > > alderson@netcom.netcom.com (Richard M. Alderson III) wrote: > > > "Jack Peacock" writes: > > > > > >>First, let's not forget DECnet goes back quite a way, mid 70's, before > > >>Ethernet even got started. > > > > > >Ethernet dates from around 1971 or 1972, as does the first > > >version of DECNET. > > > > I don't remember it being that early. What I do remember is > > There was a paper on Ethernet published by the PARC people in CACM in > July, 1976. This was probably not the first time Ethernet was ever > mentioned, but it dates it somewhere around that time, I'd say. In an paper (Personal Distributed Computing, the Alto and Ethernet Hardware) held by Charles Thacker at the 1986 ACM conference History of Personal Workstations he gives following dates: In Nov 1972 implementation of Alto begins, completed in approx 2 months. The originals had no Ethernet, but during 1973 substantial progress was made. The name Ethernet was first used in May 1973, packets being exchanged by end of 1973. -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### Message-ID: <366416E4.5BED@gazonk.del> Date: Tue, 01 Dec 1998 11:18:44 -0500 From: "Foobar T. Clown" Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 1 Dec 1998 11:19:57 -0500, 198.133.110.49 Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!nntp.abs.net!newsfeed.enteract.com!ix.netcom.com!zeus.nomos.com!198.133.110.49 jmfbahciv@aol.com wrote: > > So what is Ethernet? The local area network technology invented by David Boggs and Robert Metcalfe at Xerox corp, developed into a commercial product by Xerox, Intel, and Digital Equipment Corp, and the basis for IEEE standard, 802.3. The biggest changes to happen in all that time were the change from 3M bits per second to 10Mbps, and the changes from 1/2" heavy coax (now called 10base5) to thin coax (10base2, a.k.a., cheapernet), and finally to twisted pair (10baseT). Everything else (i.e., the OSI layer 2 protocols) has pretty much stayed the same since the beginning. There was some minor electrical difference between the original Ethernet and the 802.3 version (called "Ethernet Version II" by Xerox, Intel, and DEC). I don't remember what it was exactly, but if you mixed Version I transceivers and 802.3 transceivers on the same cable, the results were not always satisfactory. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 01 Dec 98 11:57:19 GMT Organization: UltraNet Communications, Inc. Lines: 39 Message-ID: <740mrt$eub$2@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> NNTP-Posting-Host: d3.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 1 Dec 1998 12:19:09 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!newsfeed.xcom.net!news.ultranet.com!d3 In article , Neil Franklin wrote: >Joseph Allen Dane writes: >> >> jmfbahciv@aol.com writes: >> >> > alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >> > > "Jack Peacock" writes: >> > > >> > >>First, let's not forget DECnet goes back quite a way, mid 70's, before >> > >>Ethernet even got started. >> > > >> > >Ethernet dates from around 1971 or 1972, as does the first >> > >version of DECNET. >> > >> > I don't remember it being that early. What I do remember is >> >> There was a paper on Ethernet published by the PARC people in CACM in >> July, 1976. This was probably not the first time Ethernet was ever >> mentioned, but it dates it somewhere around that time, I'd say. > >In an paper (Personal Distributed Computing, the Alto and Ethernet >Hardware) held by Charles Thacker at the 1986 ACM conference History >of Personal Workstations he gives following dates: > >In Nov 1972 implementation of Alto begins, completed in approx 2 >months. The originals had no Ethernet, but during 1973 substantial >progress was made. The name Ethernet was first used in May 1973, >packets being exchanged by end of 1973. > I guess I had better ask for a definition of Ethernet :-). So what is Ethernet? /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 01 Dec 98 11:58:17 GMT Organization: UltraNet Communications, Inc. Lines: 31 Message-ID: <740mtn$eub$3@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> NNTP-Posting-Host: d3.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 1 Dec 1998 12:20:07 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d3 In article <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu>, Joseph Allen Dane wrote: >jmfbahciv@aol.com writes: > >> In article , >> alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >> >In article <73g0uo$o0k$1@supernews.com> "Jack Peacock" >> >> >writes: >> > >> >>First, let's not forget DECnet goes back quite a way, mid 70's, before >> >>Ethernet even got started. >> > >> >Ethernet dates from around 1971 or 1972, as does the first >> >version of DECNET. >> >> I don't remember it being that early. What I do remember is > >.... > >There was a paper on Ethernet published by the PARC people in CACM in >July, 1976. This was probably not the first time Ethernet was ever >mentioned, but it dates it somewhere around that time, I'd say. > There's a _huge_ difference between a spec and implementation. DECnet people dithered for years on the spec. /BAH Subtract a hundred and four for e-mail. ###### Newsgroups: alt.folklore.computers From: kmw@ichtys.n-online.de () Subject: Re: Is VMS taboo, or what? X-Newsreader: TIN [version 1.2 PL2] Sender: usenet@ichtys.n-online.de NNTP-Posting-Host: localhost Organization: /home, sweet /home Message-ID: <1998Dec1.163051.24879@ichtys.n-online.de> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <365d9b9c.101373846@news.mcs.net> Date: Tue, 1 Dec 1998 16:30:51 GMT Lines: 21 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!newsfeed.nacamar.de!blackbush.xlink.net!noris.net!ichtys.n-online.de!ichtys.n-online.de!kmw Arargh! (arargh@arargh.com) wrote: : On Tue, 24 Nov 1998 20:36:39 +0000, "Alan J. Wylie" : wrote: : >"D. Peschel" wrote: : >[SNIP] : >> because VMS never : >> completely embraced the Internet, I think. : > : >Not just failed to embrace, but invented their own : >broken alternative to TCP/IP - DECNET : ah yes, the (*&% that changes the MAC address of the card rather than : use whats there you can do that with a number of environments.... done that some years ago in a multi-segmented 802.5 lan with LanManager being the main NOS... MAC adresses were overridden by an organisational scheme.. also sna over Netbios reuired such tricks. -- Karsten M. Winkovics eMail (home):kmw@ichtys.n-online.de eMail (Work):kmw@adtranz.de ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <36651ec0.674074252@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 20 Date: Tue, 01 Dec 1998 17:13:20 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 912532434 129.85.24.56 (Tue, 01 Dec 1998 12:13:54 EDT) NNTP-Posting-Date: Tue, 01 Dec 1998 12:13:54 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news.maxwell.syr.edu!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On Tue, 01 Dec 1998 11:18:44 -0500, "Foobar T. Clown" wrote: >jmfbahciv@aol.com wrote: >> >> So what is Ethernet? > >The local area network technology invented by David Boggs and Robert >Metcalfe at Xerox corp, developed into a commercial product by Xerox, >Intel, and Digital Equipment Corp, and the basis for IEEE standard, >802.3. [ lots of technical details snipped ] I'd say that definitive feature of Ethernet is collision-detection mechanism. Other networks use either some kind of synchronicity (like Token Ring) or re-transmission (like daisy-chain ones, now diseased). In Ethernet, IIRC, you just start transmitting, and if collision is detected, both transmitters wait a random short interval, and start again. [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### From: bhahn@spam-spam.go-away.com (Brendan Hahn) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> Organization: Transoft Corp Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit Reply-To: bhahn@transoft.mangle.net (unmangle address to reply) X-Newsreader: Yet Another NewsWatcher 2.4.0 Lines: 30 Date: Tue, 01 Dec 1998 19:01:15 GMT NNTP-Posting-Host: 205.180.87.35 NNTP-Posting-Date: Tue, 01 Dec 1998 11:01:15 PDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!207.20.0.50!peerfeed.ncal.verio.net!nntp2.cerf.net!socal.verio.net!nntp.ni.net!nnrp2.ni.net!not-for-mail alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >"Foobar T. Clown" wrote: > >>jmfbahciv@aol.com wrote: >>> >>> So what is Ethernet? >> >>The local area network technology invented by David Boggs and Robert >>Metcalfe at Xerox corp, developed into a commercial product by Xerox, >>Intel, and Digital Equipment Corp, and the basis for IEEE standard, >>802.3. >[ lots of technical details snipped ] > >I'd say that definitive feature of Ethernet is collision-detection mechanism. Ethernet is the chief example of the CSMA/CD (carrier sensing multiple access + collision detection) family of networking technologies, but it's not the only one. It may have been the first to incorporate collision detection with CSMA, or it may have been the first to use exponential backoff for collision recovery (anybody know?), but the basic features aren't unique to Ethernet. There's Fibernet, for example, a successor to Ethernet for fiberoptic connections, which has the same transmission protocol but is obviously different at the wire level. Of course the name "Ethernet", because of the success of the original, has been subjected to some loose usage--we have Fast Ethernet, gigabit Ethernet, and suchlike which are really quite distinct. Strictly speaking, it's only what Foobar described. bhahn@transoft.mangle.net <-- unmangle to reply ###### Message-ID: <366463DB.D2B@gazonk.del> Date: Tue, 01 Dec 1998 16:47:07 -0500 From: "Foobar T. Clown" Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 1 Dec 1998 16:48:20 -0500, 198.133.110.49 Lines: 28 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!dca1-hub1.news.digex.net!digex!cyclone.i1.net!newsfeed.enteract.com!ix.netcom.com!zeus.nomos.com!198.133.110.49 Alexandre Pechtchanski wrote: > I'd say that definitive feature of Ethernet is collision-detection > mechanism. Other networks use [...] If you're looking for a DEFINITION, it's IEEE 802.3. Collision detection is perhaps the most distinguishing feature, but Ethernet is not the only LAN to have used it. I don't know of another example that's current, but I know of an example from the past: ChaosNet, from MIT. ChaosNet was Ethernet with a difference... The feature that set Chaos apart was called the "virtual token" Each network interface had an eight bit counter that represented the location of the virtual token. All of these counters incremented at approximately the same rate. An interface was only allowed to transmit when it held the virtual token (i.e., when it's counter value equaled its own address), and whenever any packet was sent, all the interfaces would set their token counters to the address of the one that sent the packet. Since the counters didn't necessarily tick at exactly the same rate, and since interfaces could be powered on and off at any time, collisions were possible, and they were detected and dealt with in Ethernet-like ways. In practice, though, collsions were rare because whenever there was traffic on the net, all the token counters would be synchronized, and each interface would transmit only in its own time slot. ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> <366463DB.D2B@gazonk.del> From: Marco S Hyman Date: 01 Dec 1998 17:48:27 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 Lines: 16 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 912563308 222 marc@204.94.187.130 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news2.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail "Foobar T. Clown" writes: > Alexandre Pechtchanski wrote: > > I'd say that definitive feature of Ethernet is collision-detection > > mechanism. Other networks use [...] > > If you're looking for a DEFINITION, it's IEEE 802.3. Uhhh, 802.3 is not Ethernet. 802.3 is ethernet signaling with a different packet format. Replace the ethernet frame type with a length and then add the LLC (Logical Link Control) header and the SNAP (Sub Network Access Protocol) header to convert ethernet to 802.3 (assuming you want to run IP packets on both). Same hardware, different protocols. // marc ###### From: mwilson@interlog.com (Mel Wilson) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 2 Dec 1998 00:01:35 -0500 Organization: Interlog Internet Services -Voice (416) 975-2655 -Data 515-1414 Lines: 20 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <73f79j$rit$1@sword.avalon.net> <73g0uo$o0k$1@supernews.com> <73rfg5$dku$4@strato.ultra.net> <29k90dnj1i.fsf@honlab.nmfs.hawaii.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> NNTP-Posting-Host: shell1.interlog.com NNTP-Posting-Time: 2 Dec 1998 05:01:37 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!peerfeed.ncal.verio.net!newsfeed.interlog.com!news.interlog.com!not-for-mail In article <36651ec0.674074252@Rockyd>, alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >On Tue, 01 Dec 1998 11:18:44 -0500, "Foobar T. Clown" wrote: >>jmfbahciv@aol.com wrote: >>> So what is Ethernet? >>The local area network technology invented by David Boggs and Robert >>Metcalfe at Xerox corp, developed into a commercial product by Xerox, >>Intel, and Digital Equipment Corp, and the basis for IEEE standard, >>802.3. [ ... ] >I'd say that definitive feature of Ethernet is collision-detection mechanism. >Other networks use either some kind of synchronicity (like Token Ring) or >re-transmission (like daisy-chain ones, now diseased). In Ethernet, IIRC, you >just start transmitting, and if collision is detected, both transmitters wait a >random short interval, and start again. Is it a relative of the broadcast-radio Alohanet that the University of Hawaii was running back in the 19?70?s (which I think I vaguely remember from old ACM Computing Surveys or something?) Thanks. Mel. ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 2 Dec 1998 03:05:13 +0100 Organization: [Posted via] Interactive Networx Lines: 26 Message-ID: <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> NNTP-Posting-Host: n243-75.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newshub.northeast.verio.net!newsfeed.ecrc.net!unlisys!news.snafu.de!jnickelsen wrote: > For example paths are disk:[dir1.dir2.dir3]file.ext;n My favourite nit to pick: Full-fledged VMS pathnames were hostname"username password"::disk:[dir1.dir2.dir3]name.extension;version In most contexts for all but name.extension defaults were active, in some even for the extension. This is why the command to change the current directory is "set default" -- strange for most, but very consistent, as it sets the default for the pajor part of the pathname. "Strange, but consistent" is IMO typical for a Unix person's experiences with VMS. Of course, VMS beginners from other backgrounds may find commands like "dir" (or "directory") to list a directory or "copy" to copy files not *that* strange, perhaps not even "run" to run executable programs (necessary if they are not installed as commands). In lots of ways VMS is very consistent, and in lots of ways it is very flexible. These are, though, very different from the consistent (where existent) and flexible ways of Unix. This is the reason that Unix people very often find VMS strange and develop an intense dislike towards VMS. -- Juergen Nickelsen ###### From: bayko@pollux.cs.uregina.ca (John Bayko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 2 Dec 1998 04:03:37 GMT Organization: University of Regina, Dept. of Computer Science Lines: 28 Message-ID: <742e6p$3bc$1@sue.cc.uregina.ca> References: <73ela0$51h@hermes.acs.unt.edu> <740mrt$eub$2@strato.ultra.net> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> NNTP-Posting-Host: pollux.cs.uregina.ca Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!192.26.210.166!sunqbc.risq.qc.ca!cyclone.mbnet.mb.ca!canopus.cc.umanitoba.ca!tribune.usask.ca!news.uregina.ca!not-for-mail In article <36651ec0.674074252@Rockyd>, Alexandre Pechtchanski wrote: >On Tue, 01 Dec 1998 11:18:44 -0500, "Foobar T. Clown" wrote: >>jmfbahciv@aol.com wrote: >>> >>> So what is Ethernet? >> >>The local area network technology invented by David Boggs and Robert >>Metcalfe at Xerox corp, developed into a commercial product by Xerox, >>Intel, and Digital Equipment Corp, and the basis for IEEE standard, >>802.3. >[ lots of technical details snipped ] > >I'd say that definitive feature of Ethernet is collision-detection mechanism. >Other networks use either some kind of synchronicity (like Token Ring) or >re-transmission (like daisy-chain ones, now diseased). In Ethernet, IIRC, you >just start transmitting, and if collision is detected, both transmitters wait a >random short interval, and start again. I've often wondered why people preferred a technology where usable bandwidth actually *decreased* as network traffic increased, over a technology like token ring. Presumably, like VHS and Windows, because it's 'good enough' despite being not perfect. -- John Bayko (Tau). bayko@cs.uregina.ca http://www.cs.uregina.ca/~bayko ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 02 Dec 1998 06:17:02 GMT Organization: Arargh!! Lines: 29 Message-ID: <3665d736.163078449@news.mcs.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <365d9b9c.101373846@news.mcs.net> <1998Dec1.163051.24879@ichtys.n-online.de> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!nntp.abs.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On Tue, 1 Dec 1998 16:30:51 GMT, kmw@ichtys.n-online.de () wrote: >Arargh! (arargh@arargh.com) wrote: >: On Tue, 24 Nov 1998 20:36:39 +0000, "Alan J. Wylie" >: wrote: > >: >"D. Peschel" wrote: >: >[SNIP] >: >> because VMS never >: >> completely embraced the Internet, I think. >: > >: >Not just failed to embrace, but invented their own >: >broken alternative to TCP/IP - DECNET >: ah yes, the (*&% that changes the MAC address of the card rather than >: use whats there > >you can do that with a number of environments.... done that some years ago >in a multi-segmented 802.5 lan with LanManager being the main NOS... MAC >adresses were overridden by an organisational scheme.. also sna over Netbios that the way DECnet works(worked?), it changed the mac addr to something like 0xaa00 + the DECnet Node address. The problem that I see is what happens if you need to merge two existing independent DECnet lans that just happen to have the same base address? This just seems to be a bad design arargh >reuired such tricks. ###### From: Joseph Allen Dane Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 02 Dec 1998 08:27:50 -1000 Organization: University of Hawaii Lines: 17 Sender: jdane@honlab.nmfs.hawaii.edu Message-ID: <29btlmmmah.fsf@honlab.nmfs.hawaii.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> NNTP-Posting-Host: honlab.nmfs.hawaii.edu Mime-Version: 1.0 (generated by tm-edit 7.106) Content-Type: text/plain; charset=US-ASCII X-Newsreader: Gnus v5.6.9/Emacs 19.33 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!news.he.net!news.lava.net!news.Hawaii.Edu!news sarr@engin.umich.edu (Sarr J. Blumson) writes: > > Is it a relative of the broadcast-radio Alohanet that the University > >of Hawaii was running back in the 19?70?s (which I think I vaguely > >remember from old ACM Computing Surveys or something?) > > Yes. In fact if you read the CACM paper by Boggs and Metcalfe (from 197?) > they _talk_ about Alohanet as the source for the idea. > CAMC, July 1976 Volume 19, n.7 "Ethernet: Distributed Packet Switching for Local Computer Networks" -- joe ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 2 Dec 1998 09:17:39 GMT Organization: Daedalus Consulting Lines: 34 Message-ID: <7430jj$6g0$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <36651ec0.674074252@Rockyd> <366463DB.D2B@gazonk.del> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article , Marco S Hyman wrote: >Uhhh, 802.3 is not Ethernet. 802.3 is ethernet signaling with >a different packet format. Replace the ethernet frame type >with a length and then add the LLC (Logical Link Control) header >and the SNAP (Sub Network Access Protocol) header to convert >ethernet to 802.3 (assuming you want to run IP packets on both). >Same hardware, different protocols. Close, but no cigar. IEEE 802.3 (aka ISO 8802.3) and its annexes define the hardware from old fashioned big yellow cable (etherhose 8-) through to 100 Mbps (and gigabit?) over fibre & twisted pair. It does assume that 802.2 packet formats (LLC/SNAP) will be used however. But 802.3 hardware will happily use the DIX Ethernet II packet format -- the difference being that the DIX format goes: preamble sfd destination source protocol data crc 63bits 1b 48bits 48bits 16bits 16b wherease 802.2/802.3 (ie 802.2 headers with 802.3 framing) goes preamble sfd destination source llc/snap data crc 63bits 1b 48bits 48bits variable 16b and the LLC stuff starts with a packet length. But by a lucky accident 8-) the DIX style protocol numbers (when converted to network byte order) are all greater than the maximum Ethernet packet size, so they never conflict. There are some other subtle differences between Ethernet II and 802.3, mainly relating to repeaters. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> Reply-To: sarr@umich.edu Organization: University of Michigan Subject: Re: Is VMS taboo, or what? From: sarr@engin.umich.edu (Sarr J. Blumson) Lines: 26 Message-ID: Date: Wed, 02 Dec 1998 14:32:48 GMT NNTP-Posting-Host: 141.211.43.166 X-Trace: news.itd.umich.edu 912609168 141.211.43.166 (Wed, 02 Dec 1998 09:32:48 EDT) NNTP-Posting-Date: Wed, 02 Dec 1998 09:32:48 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!141.211.144.13!newsxfer3.itd.umich.edu!news.itd.umich.edu!sarr In article , Mel Wilson wrote: >In article <36651ec0.674074252@Rockyd>, >alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >>On Tue, 01 Dec 1998 11:18:44 -0500, "Foobar T. Clown" wrote: >>>jmfbahciv@aol.com wrote: >>>> So what is Ethernet? >>>The local area network technology invented by David Boggs and Robert >>>Metcalfe at Xerox corp, developed into a commercial product by Xerox, >>>Intel, and Digital Equipment Corp, and the basis for IEEE standard, >>>802.3. [ ... ] >>I'd say that definitive feature of Ethernet is collision-detection mechanism. ... > Is it a relative of the broadcast-radio Alohanet that the University >of Hawaii was running back in the 19?70?s (which I think I vaguely >remember from old ACM Computing Surveys or something?) Yes. In fact if you read the CACM paper by Boggs and Metcalfe (from 197?) they _talk_ about Alohanet as the source for the idea. -- -------- Sarr Blumson sarr@umich.edu voice: +1 734 764 0253 home: +1 734 665 9591 ITD, University of Michigan http://www-personal.umich.edu/~sarr/ 519 W William, Ann Arbor, MI 48103-4943 ###### From: atbowler@thinkage.on.ca (Alan Bowler) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 2 Dec 1998 18:40:18 GMT Organization: Thinkage Ltd. Lines: 17 Message-ID: <7441ii$1qj$1@nntp2.uunet.ca> References: <366416E4.5BED@gazonk.del> <36651ec0.674074252@Rockyd> <742e6p$3bc$1@sue.cc.uregina.ca> NNTP-Posting-Host: 192.102.11.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.direct.ca!news.insinc.net!news1.van.metronet.ca!news1.cal.metronet.ca!news.uunet.ca!atbowler In article <742e6p$3bc$1@sue.cc.uregina.ca> bayko@pollux.cs.uregina.ca (John Bayko) writes: > > I've often wondered why people preferred a technology where usable >bandwidth actually *decreased* as network traffic increased, over a >technology like token ring. Presumably, like VHS and Windows, because >it's 'good enough' despite being not perfect. I used to wonder about Ethernet for the same reason, until I saw and article in CACM (no I don't remember exactly when) that had a study of how well Ethernet's actually worked under excessive overload. It seems that it flattens out nicely, and you end with it delivering about 85% of its capacity in successful transfers. (i.e. about 15% overhead). This is comparable to the bookeeping overheads for some other protocols, and is as you say "good enough". It may be a reasonable price to pay given Ethernet's superior reliability because of its likely failure modes. (i.e. experience and theory show Ethernet to more robust within its applicable domain). ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 02 Dec 1998 19:34:05 GMT Organization: Arargh!! Lines: 23 Message-ID: <36659552.38143011@news.mcs.net> References: <73ela0$51h@hermes.acs.unt.edu> <36651ec0.674074252@Rockyd> <366463DB.D2B@gazonk.del> <7430jj$6g0$1@news.wlg.netlink.net.nz> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!nntp.abs.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On 2 Dec 1998 09:17:39 GMT, don@news.daedalus.co.nz (Don Stokes) wrote: >In article , >Marco S Hyman wrote: >and the LLC stuff starts with a packet length. But by a lucky accident >8-) the DIX style protocol numbers (when converted to network byte >order) are all greater than the maximum Ethernet packet size, so they >never conflict. Mostly, but not quite, I quote from RFC1700 starting at line 9380: Ethernet Exp. Ethernet Description References ------------- ------------- ----------- ---------- decimal Hex decimal octal 000 0000-05DC - - IEEE802.3 Length Field [XEROX] 257 0101-01FF - - Experimental [XEROX] 512 0200 512 1000 XEROX PUP (see 0A00) [8,XEROX] 513 0201 - - PUP Addr Trans (see 0A01)[XEROX] 0400 Nixdorf [XEROX] a couple of odd ones snuck in arargh ###### From: "The XO" Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> Subject: Re: Is VMS taboo, or what? Date: Thu, 3 Dec 1998 00:02:00 +1030 Lines: 56 Organization: Headquarters X-Newsreader: Microsoft Outlook Express 4.71.1712.3 X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3 NNTP-Posting-Host: netcafe.pirie.mtx.net.au Message-ID: <366542cc.0@news.mtx.net.au> X-Trace: 3 Dec 1998 00:08:20 -1050, netcafe.pirie.mtx.net.au Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!news.maxwell.syr.edu!news.mel.connect.com.au!news.syd.connect.com.au!nsw.nntp.telstra.net!news1.mpx.com.au!news1.optus.net.au!optus!yorrell.saard.net!duster.adelaide.on.net!news.saix.saia.asn.au!news.mtx.net.au!netcafe.pirie.mtx.net.au lee1089@kettering.edu wrote in message ... >[1] I once saw a VMS server run completely out of disk space. Not one >single block was left. However, the server continued to run (albeit >somewhat slowly) and once some space was freed on the hard drive it >ran normally again. I must say that I was impressed but not enough to >overcome my dislike of the command syntax. I can beat that. I have seen a VMS system, on the internet, in the midst of an ftp download, have its HSC (Fancy disk controller - PDP11 based) knocked out by a power glitch. The Vax had an internal UPS and kept going, VMS noted that it could no longer talk to ANY of it's drives or tapes, (including the one with it's swap and page files!) and started MOUNT/VERIFY to try and get them back, polling the HSC for a response. (The VAX and HSC are connected by a thick cable 70mbs system called CI Bus - Computer Interconnect). Since the HSC boots it's little O/S (CRONIC) off TU58 cassette tape drives, it takes about 8 minutes to reboot and come back online. When it finally did, VMS happily remounted the drives, there was a flurry of disk activity while it got caught up, then it carried on as normal. The FTP transfer? Oh, yeah, it kept going, and had actually finished transferring the data from the remote host BEFORE the disks came back on line, so it buffered it, waited for the drives to come back, then wrote it to the disk. Loss of data 0, Loss of function, minimal. Coolness factor, very high. I was impressed anyway. Anyone want to try unplugging the hard drive on a Windows NT server?.......No?....Hmm, wonder why? I'll live with the command line, since I grew up with dos, not windoze. Unix commands can get just as complicated, and the some of the commands are even more cryptic than the vms equivalent. BIFF is named after a dog for heavens sake. BTW, there is no equivalent BIFF daemon under VMS, the function is BUILT IN to the O/S. I have no real problem with either, since both are fine examples of real operating systems, Windoze NT/95/98/2000 whatever being fine examples of a twisted attempt at an operating system. (IMHO) Cheers Geoff Computer Room Internet Cafe Port Pirie South Australia Spam countermeasures netcafe at pirie dot mtx dot net dot au ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 3 Dec 1998 01:01:15 GMT Organization: La Trobe University Lines: 27 Message-ID: <744nsr$4ll$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 912646875 4789 131.172.12.11 (3 Dec 1998 01:01:15 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 3 Dec 1998 01:01:15 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Juergen Nickelsen (jnickelsen@acm.org) wrote: : "Strange, but consistent" is IMO typical for a Unix person's experiences : with VMS. Of course, VMS beginners from other backgrounds may find : commands like "dir" (or "directory") to list a directory or "copy" to : copy files not *that* strange, perhaps not even "run" to run executable : programs (necessary if they are not installed as commands). Of course those of us who had to manage VMS and AOS systems had lots of fun with the directory command.... : In lots of ways VMS is very consistent, and in lots of ways it is very : flexible. These are, though, very different from the consistent (where : existent) and flexible ways of Unix. This is the reason that Unix people : very often find VMS strange and develop an intense dislike towards VMS. Remember, a lot of what people define as the difference between VMS and Unix is really a discussion about the difference between DCL and their favorite shell. There's a lot more to consider than syntactic sugar. In the days when Posix was a supported product under VMS you could really confuse users by starting a Unix like environment. (Sort of reminds me about Eunice but every time I think about that I feel unwell :-) -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 3 Dec 1998 01:03:40 GMT Organization: La Trobe University Lines: 14 Message-ID: <744o1c$4ll$2@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <365d9b9c.101373846@news.mcs.net> <1998Dec1.163051.24879@ichtys.n-online.de> <3665d736.163078449@news.mcs.net> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 912647020 4789 131.172.12.11 (3 Dec 1998 01:03:40 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 3 Dec 1998 01:03:40 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!newsfeed.wli.net!su-news-hub1.bbnplanet.com!su-news-feed4.bbnplanet.com!news.gtei.net!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Arargh! (arargh@arargh.com) wrote: : that the way DECnet works(worked?), it changed the mac addr to : something like 0xaa00 + the DECnet Node address. The problem that I : see is what happens if you need to merge two existing independent : DECnet lans that just happen to have the same base address? This just : seems to be a bad design Well TCP/IP fares no better should you merge two LANS that have computers with the same IP address. -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: dpeschel@u.washington.edu (D. Peschel) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 3 Dec 1998 05:27:20 GMT Organization: University of Washington, Seattle Lines: 39 Message-ID: <7457fo$16em$1@nntp1.u.washington.edu> References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <744nsr$4ll$1@news.latrobe.edu.au> NNTP-Posting-Host: saul7.u.washington.edu X-Trace: nntp1.u.washington.edu 912662840 39382 (None) 140.142.17.39 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: dpeschel Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!feed1.news.luth.se!luth.se!sunqbc.risq.qc.ca!newsfeed.cwix.com!192.220.250.21!netnews1.nw.verio.net!netnews.nwnet.net!news.u.washington.edu!dpeschel In article <744nsr$4ll$1@news.latrobe.edu.au>, Huw Davies wrote: >Juergen Nickelsen (jnickelsen@acm.org) wrote: >: "Strange, but consistent" is IMO typical for a Unix person's experiences >: with VMS. Of course, VMS beginners from other backgrounds may find >: commands like "dir" (or "directory") to list a directory or "copy" to >: copy files not *that* strange, perhaps not even "run" to run executable >: programs (necessary if they are not installed as commands). On the surface the commands make sense, but when you start looking at the details, you find many differences from UNIX. For example, SET DEFAULT will let you set a default directory that doesn't actually exist. And DIR has a ton of options, because of the complexity of the VMS filesystem. >Remember, a lot of what people define as the difference between VMS and >Unix is really a discussion about the difference between DCL and their >favorite shell. There's a lot more to consider than syntactic sugar. In the >days when Posix was a supported product under VMS you could really confuse >users by starting a Unix like environment. (Sort of reminds me about >Eunice but every time I think about that I feel unwell :-) What about the difference between a UNIX shell and MCR? You can't leave MCR out of this. :) I always got the impression that replacing DCL was much harder than replacing a UNIX shell would be. Is this true, or not? At any rate, the VMS systems I've used never offer any other options besides DCL. While installing Perl 4 (on a UNIX system) the configuration script gave me the following message: Congratulations, you're not using Eunice. Something like that, anyway. -- Derek ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 3 Dec 1998 09:39:08 GMT Organization: Daedalus Consulting Lines: 80 Message-ID: <745m7s$g4m$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <744nsr$4ll$1@news.latrobe.edu.au> <7457fo$16em$1@nntp1.u.washington.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <7457fo$16em$1@nntp1.u.washington.edu>, D. Peschel wrote: >On the surface the commands make sense, but when you start looking at the >details, you find many differences from UNIX. > >For example, SET DEFAULT will let you set a default directory that doesn't >actually exist. And your point? $ set default [foo.bar.baz] ! None of these exist $ create/dir [] as opposed to: % mkdir /foo % mkdir /foo/bar % mkdir /foo/bar/baz % cd /foo/bar/baz >And DIR has a ton of options, because of the complexity of the VMS filesystem. Hardly. "ls" on my BSD system has 22 options. DIRECTORY on my VMS box has 35. But then, DIRECTORY lets you specify individually which fields to display, rather than just giving you a small handful of possible output formats. It also has several selection criteria, such as listing files of certain sizes or ages, functions that "ls" doesn't have. Bizarrely, it even has an option to do a directory listing via FTP. There are a couple of options to deal with features that most unices simply don't have, such as file version numbers and ACLs, but very few. If anything, the much more general file/directory parsing rules of VMS reduce the number of options -- there's no equivalent to -R for example, because you can specify "..." in the directory spec to do a recursive listing (and that's supported right down for a fairly low level -- none of this opening a directory file yourself stuff...). >I always got the impression that replacing DCL was much harder than replacing >a UNIX shell would be. Is this true, or not? At any rate, the VMS systems >I've used never offer any other options besides DCL. There are a bunch of minor services that are "implemented by the CLI". That means they're supervisor-mode callbacks into DCL, and do things like set DCL symbols, spawn CLI suprocesses and set supervisor mode logical names. Several of the LIB$ library calls use these, and don't work if the CLI (ie DCL or MCR) is not mapped in. And that's on top of the CLI stuff. There is a version of tcsh floating around for use with the POSIX stuff, and that's on top of the "POSIX CLI" which vaguely resembles the Bourne shell. But for executing DCL commands, I believe these both spawn subprocesses with DCL mapped in. DCL is ugly. It tends to take more DCL than unix shell code to do anything, although DCL can quite often do stuff internally that a unix shell would need to launch external programs for. Mainly, it would be nice of I/O redirection could be done in the context of a command rather than as separate ASSIGN statements, and if the control structures were a bit better than just GOTO and GOSUB (and don't get me started on CALL). On the other hand, unix shells IMAO would be vastly improved if they had DCL's style lexical functions. The command structure isn't bad; it takes a bit of getting used to the fact that commands and programs aren't the same thing as they are in unix, DOS and various other OSes. While it's not quite so quick to type "/FILE_ID" as "-i", command scripts in DCL tend to be a hell of a lot more readable... DCL might be ugly on the outside, but it's an absolute nightmare on the inside. I'm pretty sure the reason it doesn't get much work done on it to improve it is simply that nobody is brave enough to go near the code. I'm still waiting for someone who knows the Alpha better than I do to go and port DCLCOMPLETE (which by a disgusting, rather VAX-specific hack that I'm not particularly proud of adds command and filename completion to DCL) to the Alpha, but I shudder to think what the ported DCL code looks like inside... -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: tph@longhorn.uucp (Tom Harrington) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 3 Dec 1998 17:20:44 GMT Organization: Mechanist Industries Lines: 16 Message-ID: <746h9c$j722@eccws1.dearborn.ford.com> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> Reply-To: tph@rmi.net NNTP-Posting-Host: cs0053.eld.ford.com X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.nyu.edu!newsxfer3.itd.umich.edu!jobone!dailyplanet.srl.ford.com!eccws1.dearborn.ford.com!longhorn!tph The XO (the_xo@yahoo.com.au) wrote: : there is no equivalent : BIFF daemon under VMS, the function is BUILT IN to the O/S. : I have no real problem with either, since both are fine examples of real : operating systems, Windoze NT/95/98/2000 whatever being fine examples of a : twisted attempt at an operating system. (IMHO) Isn't NT the bastard child of VMS? -- Tom Harrington --------- tph@rmii.com -------- http://rainbow.rmii.com/~tph "UN-altered REPRODUCTION and DISSEMINATION of this IMPORTANT information is ENCOURAGED" -Robert McElwaine Cookie's Revenge: ftp://ftp.rmi.net/pub2/tph/cookie/cookies-revenge.sit.hqx ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 03 Dec 1998 21:27:50 +0100 Organization: My own Private Self Lines: 58 Sender: neil@chonsp.franklin.ch Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <744nsr$4ll$1@news.latrobe.edu.au> <7457fo$16em$1@nntp1.u.washington.edu> <745m7s$g4m$1@news.wlg.netlink.net.nz> X-Newsreader: Gnus v5.3/Emacs 19.34 don@news.daedalus.co.nz (Don Stokes) writes: > > D. Peschel wrote: > > > >And DIR has a ton of options, because of the complexity of the VMS filesystem . > > Hardly. "ls" on my BSD system has 22 options. DIRECTORY on my VMS box > has 35. And GNU "ls" on this Linux has: SYNOPSIS ls [-abcdfgiklmnpqrstuxABCFGLNQRSUX1] [-w cols] [-T cols] [-I pattern] [--all] [--escape] [--directory] [--inode] [--kilobytes] [--numeric-uid-gid] [--no-group] [--hide- control-chars] [--reverse] [--size] [--width=cols] [--tab­ size=cols] [--almost-all] [--ignore-backups] [--classify] [--file-type] [--full-time] [--ignore=pattern] [--derefer­ ence] [--literal] [--quote-name] [--recursive] [--sort={none,time,size,extension}] [--format={long,ver­ bose,commas,across,vertical,single-column}] [--time={atime,access,use,ctime,status}] [--help] [--ver­ sion] [--color[={none,auto,always}]] [--colour[={none,auto,always}]] [name...] Thats from the man page. I am not going to count them. > But then, DIRECTORY lets you specify individually which fields > to display, rather than just giving you a small handful of possible output > formats. And is has DIR /FULL :-). About 10..15 lines per file. > Bizarrely, > it even has an option to do a directory listing via FTP. Not to forget the DIR HOST::DISK:[DIR.DIR]FILE.EXT;V syntax for DECnet visible files. Unix had to wait for URLs to get something equivalent (and that also only while in an brower). > The command structure isn't bad; it takes a bit of getting used to the > fact that commands and programs aren't the same thing as they are in > unix, DOS and various other OSes. While it's not quite so quick to type > "/FILE_ID" as "-i", command scripts in DCL tend to be a hell of a lot > more readable... ls --inode, at least with GNU ls, see above. Neil "when will I find time to set up my 'new' VAXstation" Franklin -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 03 Dec 1998 20:28:53 +0000 Organization: Tanglewood Message-ID: <3666F485.D0CA687@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912717255 nnrp-11:9321 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 28 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail The XO wrote: > then wrote it to the disk. Loss of data 0, Loss of function, minimal. > Coolness factor, very high. > I was impressed anyway. Well, having grown up with DEC kit, and once been a VAX sysadmin, I can truthfully say that I expect no less from a real computer. However it *is* impressive when it happens... > I have no real problem with either, since both are fine examples of real > operating systems, Windoze NT/95/98/2000 whatever being fine examples of a > twisted attempt at an operating system. (IMHO) Since I can't run VMS on this PC I am running Linux, which comes a close second to VMS in my estimation. I would still like to see what would happen if someone returned a copy of NT to M$ and demanded a refund on the grounds that it was "not of merchentable quality", citing the behaviour of VMS as the standard. Certainly this would be possible under UK law, where discalimers are of little value, and anything *inside* the shrink-wrap has no force at all. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 03 Dec 1998 20:29:58 +0000 Organization: Tanglewood Message-ID: <3666F4C6.362324BF@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <746h9c$j722@eccws1.dearborn.ford.com> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912717496 nnrp-10:7113 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 11 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Tom Harrington wrote: > Isn't NT the bastard child of VMS? Yes, but we don't mention it, mother gets *so* upset... -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 03 Dec 1998 21:20:29 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 11 Message-ID: <3667003e.1041305@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.pop.de!newsfeed.germany.net!wuff.mayn.de!unlisys!news.snafu.de!home.ibert.com!news On Wed, 2 Dec 1998 03:05:13 +0100, jnickelsen@acm.org (Juergen Nickelsen) wrote: >In lots of ways VMS is very consistent, and in lots of ways it is very >flexible. These are, though, very different from the consistent (where >existent) and flexible ways of Unix. This is the reason that Unix people >very often find VMS strange and develop an intense dislike towards VMS. Actually, I developed a deep-felt dislike for VMS before I actually fell in love with UNIX. Maybe I was a UNIX guy before I myself knew it. ###### From: ejb@klamath.demon.co.uk (Edward John M. Brocklesby) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 03 Dec 1998 21:52:39 GMT Message-ID: <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> NNTP-Posting-Host: klamath.demon.co.uk X-NNTP-Posting-Host: klamath.demon.co.uk:212.228.198.242 X-Trace: news.demon.co.uk 912721959 nnrp-07:24008 NO-IDENT klamath.demon.co.uk:212.228.198.242 X-Complaints-To: abuse@demon.net X-Newsreader: slrn (0.9.4.3 UNIX) Lines: 30 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.ecrc.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!klamath.demon.co.uk!ejb On about Thu, 3 Dec 1998 00:02:00 +1030, The XO might have written: [snip story about HD] I have a Linux box here that has a single 8.5Gb EIDE HD running the whole thing. Occasionally, I do something stu^H^H^Hinteresting, like drop it, or turn it off accidently (it's on a seperate PSU to the computer), and whatever I do, Linux manages to recover from it fine. (Unlike NetBSD, which just hangs when I turn the HD off). That's on Amiga hardware, don't know if it would work on PC's. >Cheers > >Geoff >Computer Room Internet Cafe >Port Pirie >South Australia >Spam countermeasures >netcafe at pirie dot mtx dot net dot au > > -- Edward John M. Brocklesby System Administrator, Klamath Public Access Unix System | Data: +44 1865 454802 | Free-VMS - Expand your mind - www.free-vms.org | Free UNIX Accounts | Finger ejb@deep-thought.ml.org for PGP key ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 01:00:25 GMT Organization: La Trobe University Lines: 19 Message-ID: <747c79$jo7$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 912733225 20231 131.172.12.11 (4 Dec 1998 01:00:25 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 4 Dec 1998 01:00:25 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd The XO (the_xo@yahoo.com.au) wrote: : I can beat that. I have seen a VMS system, on the internet, in the midst : of an ftp download, : have its HSC (Fancy disk controller - PDP11 based) knocked out by a power : glitch. From memory, this functionality was added to VMS v3 and applies to any VMS mounted disk. I remember testing this after the VMS upgrade. Even though the manual said that it would be OK, it took some guts to spin down the system disk (an RM80) with the system running.... This must have been around 1983. I guess other operating systems have a fair way to go.... -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 01:03:35 GMT Organization: La Trobe University Lines: 23 Message-ID: <747cd7$jo7$2@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <744nsr$4ll$1@news.latrobe.edu.au> <7457fo$16em$1@nntp1.u.washington.edu> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 912733415 20231 131.172.12.11 (4 Dec 1998 01:03:35 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 4 Dec 1998 01:03:35 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd D. Peschel (dpeschel@u.washington.edu) wrote: : I always got the impression that replacing DCL was much harder than replacing : a UNIX shell would be. Is this true, or not? At any rate, the VMS systems : I've used never offer any other options besides DCL. Well it's not hard. Write your own command interpretor and in your login.com run this. Of course writing your own command interpretor may not be trivial. For a course being run by our history department I wrote some programs that made the user environment look an aweful lot like DOS (in fact most users thought they were using DOS on the nice VT terminals...). : While installing Perl 4 (on a UNIX system) the configuration script gave me : the following message: : Congratulations, you're not using Eunice. That's exactly the one... -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 1998 05:49:04 GMT Organization: Posted via Supernews, http://www.supernews.com - Discussions start here! Lines: 29 Message-ID: <36677771.23568198@news.vip.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.32 X-Trace: 912750263 A01OARAUVD420CCD1C usenet76.supernews.com X-Complaints-To: newsabuse@supernews.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net!xara.net!news.itg.net.uk!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >The XO (the_xo@yahoo.com.au) wrote: > >: I can beat that. I have seen a VMS system, on the internet, in the midst >: of an ftp download, >: have its HSC (Fancy disk controller - PDP11 based) knocked out by a power >: glitch. > >From memory, this functionality was added to VMS v3 and applies to any >VMS mounted disk. I remember testing this after the VMS upgrade. Even though >the manual said that it would be OK, it took some guts to spin down the >system disk (an RM80) with the system running.... This is impressive. Do you have any details on how it was done? The post before said that VMS managed even with the VM/paging disk out of action. How? >This must have been around 1983. I guess other operating systems have >a fair way to go.... Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: arargh@arargh.com (Arargh!) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 1998 06:49:38 GMT Organization: Arargh!! Lines: 22 Message-ID: <36678551.2510824@news.mcs.net> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <365d9b9c.101373846@news.mcs.net> <1998Dec1.163051.24879@ichtys.n-online.de> <3665d736.163078449@news.mcs.net> <744o1c$4ll$2@news.latrobe.edu.au> Reply-To: arargh@arargh.com NNTP-Posting-Host: jwright.pr.mcs.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer.gip.net!news.gsl.net!gip.net!newspump.sol.net!news.mcs.net!ddsw1!news.mcs.net!not-for-mail On 3 Dec 1998 01:03:40 GMT, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >Arargh! (arargh@arargh.com) wrote: >: that the way DECnet works(worked?), it changed the mac addr to >: something like 0xaa00 + the DECnet Node address. The problem that I >: see is what happens if you need to merge two existing independent >: DECnet lans that just happen to have the same base address? This just >: seems to be a bad design > >Well TCP/IP fares no better should you merge two LANS that have computers >with the same IP address. True. But, maybe, you could stick a firewall between them, and let it translate. maybe Arargh ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 98 11:38:10 GMT Organization: UltraNet Communications, Inc. Lines: 33 Message-ID: <748ist$lfi$1@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <3666F485.D0CA687@tnglwood.demon.co.uk> NNTP-Posting-Host: d7.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 4 Dec 1998 12:00:29 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!feed1.news.rcn.net!rcn!news.ultranet.com!d7 In article <3666F485.D0CA687@tnglwood.demon.co.uk>, Robert Billing wrote: >The XO wrote: > >> then wrote it to the disk. Loss of data 0, Loss of function, minimal. >> Coolness factor, very high. >> I was impressed anyway. > > Well, having grown up with DEC kit, and once been a VAX sysadmin, I can >truthfully say that I expect no less from a real computer. However it >*is* impressive when it happens... > >> I have no real problem with either, since both are fine examples of real >> operating systems, Windoze NT/95/98/2000 whatever being fine examples of a >> twisted attempt at an operating system. (IMHO) > > Since I can't run VMS on this PC I am running Linux, which comes a >close second to VMS in my estimation. > > I would still like to see what would happen if someone returned a copy >of NT to M$ and demanded a refund on the grounds that it was "not of >merchentable quality", citing the behaviour of VMS as the standard. >Certainly this would be possible under UK law, where discalimers are of >little value, and anything *inside* the shrink-wrap has no force at all. > Really? My interpretation of those disclaimers was WYSIWYG, promising nothing about performance. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 98 11:42:08 GMT Organization: UltraNet Communications, Inc. Lines: 39 Message-ID: <748j4a$lfi$2@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> NNTP-Posting-Host: d7.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 4 Dec 1998 12:04:26 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newspeer.monmouth.com!news.ultranet.com!d7 In article <36677771.23568198@news.vip.net>, genew@vip.net (Gene Wirchenko) wrote: >cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: > >>The XO (the_xo@yahoo.com.au) wrote: >> >>: I can beat that. I have seen a VMS system, on the internet, in the midst >>: of an ftp download, >>: have its HSC (Fancy disk controller - PDP11 based) knocked out by a power >>: glitch. >> >>From memory, this functionality was added to VMS v3 and applies to any >>VMS mounted disk. I remember testing this after the VMS upgrade. Even though >>the manual said that it would be OK, it took some guts to spin down the >>system disk (an RM80) with the system running.... > > This is impressive. Do you have any details on how it was done? >The post before said that VMS managed even with the VM/paging disk out >of action. How? Sigh! It's called timesharing. That's an aspect of implementation of timesharing. The software needs to be robust, no matter what those pesky users do. Hardware off-line interrupts are a piece of cake in comparison. TOPS10 knew how to do all that in the 70s (if not the 60s). > >>This must have been around 1983. I guess other operating systems have >>a fair way to go.... One of my frustrations is seeing our wheels reinvented as if they had never existed. /BAH Subtract a hundred and four for e-mail. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 1998 12:56:44 +0000 Organization: Tanglewood Message-ID: <3667DC0C.F4A340B8@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <3666F485.D0CA687@tnglwood.demon.co.uk> <748ist$lfi$1@strato.ultra.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912776713 nnrp-09:27357 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail jmfbahciv@aol.com wrote: > Really? My interpretation of those disclaimers was WYSIWYG, promising > nothing about performance. OTOH if you describe something as an operating system, then it's jolly well got to be one, or you can be prosecuted under the TDA. Similarly, if inside the shrink wrap it says, "you aren't allowed to make backups," then it's too late, you can make as many backups as you like. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 98 13:05:49 GMT Organization: UltraNet Communications, Inc. Lines: 24 Message-ID: <748o17$v01$1@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <3666F485.D0CA687@tnglwood.demon.co.uk> <748ist$lfi$1@strato.ultra.net> <3667DC0C.F4A340B8@tnglwood.demon.co.uk> NNTP-Posting-Host: d8.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 4 Dec 1998 13:28:07 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newspeer.monmouth.com!news.ultranet.com!d8 In article <3667DC0C.F4A340B8@tnglwood.demon.co.uk>, Robert Billing wrote: >jmfbahciv@aol.com wrote: > >> Really? My interpretation of those disclaimers was WYSIWYG, promising >> nothing about performance. > > OTOH if you describe something as an operating system, then it's jolly >well got to be one, or you can be prosecuted under the TDA. It only has to function as an OS. It still says nothing about performance quality, e.g. not crashing a lot. > Similarly, >if inside the shrink wrap it says, "you aren't allowed to make backups," >then it's too late, you can make as many backups as you like. > Backups of what? The system disk? Hmmm...there's really bad implications there. /BAH Subtract a hundred and four for e-mail. ###### From: plj@byron.ext.telia.se (Peter Ljungberg) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 16:19:17 GMT Organization: Telia IT-Service / Torsas Computer Society http://byron.ext.telia.se Lines: 37 Message-ID: <749225$1kv$1@news.han.telia.se> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <365B18D7.20A5AC1@zen.co.uk> <365d9b9c.101373846@news.mcs <36678551.2510824@news.mcs.net> NNTP-Posting-Host: pc238.southeast.telia.se X-Newsreader: WinVN NT 0.92.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!newspost.telia.com!news.han.telia.se!not-for-mail In article <36678551.2510824@news.mcs.net>, arargh@arargh.com (Arargh!) says: > >On 3 Dec 1998 01:03:40 GMT, cchd@lucifer.its.latrobe.edu.au (Huw >Davies) wrote: > >>Arargh! (arargh@arargh.com) wrote: >>: that the way DECnet works(worked?), it changed the mac addr to >>: something like 0xaa00 + the DECnet Node address. The problem that I >>: see is what happens if you need to merge two existing independent >>: DECnet lans that just happen to have the same base address? This just >>: seems to be a bad design >> >>Well TCP/IP fares no better should you merge two LANS that have computers >>with the same IP address. > >True. > >But, maybe, you could stick a firewall between them, and let it >translate. > >maybe > >Arargh > I once by mistake, pulled the SDI-cable (black cable between computer and/or disk/controller) to the systemdisk, heard beeping from the console, which said that DUA0: (the systemdisk) was offline, so I placed it back where it was before and everything was as before.... VMS unbeatable! >>> ^P.Lj ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 04 Dec 1998 18:19:48 GMT Organization: . Lines: 51 Message-ID: <36691be2.2492303@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> NNTP-Posting-Host: pool02b-194-7-146-74.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to alt.folklore.computers: > In article <36677771.23568198@news.vip.net>, > genew@vip.net (Gene Wirchenko) wrote: > > This is impressive. Do you have any details on how it was done? > >The post before said that VMS managed even with the VM/paging disk out > >of action. How? > > Sigh! It's called timesharing. That's an aspect of implementation > of timesharing. The software needs to be robust, no matter what > those pesky users do. Hardware off-line interrupts are a piece of > cake in comparison. TOPS10 knew how to do all that in the 70s (if > not the 60s). Sorry if this is a stupid remark, but I don't see in what way the fact that it's a timesharing system can have anything to do with being able to continue whatever it's doing when the system disk goes off-line (or after it comes back on)? For all I know, timesharing is just another way to implement multitasking (in fact, resembling what we call preemptive multitasking today, in that it slices up available CPU time over different users or processes). I guess I'm about 5 years too young to really know timesharing though. BTW, some of my first steps were made on RSX-11 - and I seem to remember (vaguely, after 20 years) entering the building one morning, finding both harddisks (thus including the system disk) powered off: after switching them on, the system ran normally. And I wouldn't even find it very strange: AFAIK, the system did nothing in the background like today's systems do (in a way - sorry for the comparison - it was more like DOS rather than Windows in that aspect, at least that's how it appeared to be on that particular system: I never saw any disk activity when I wasn't running anything). Not quite the same as being able to survive power cycling in the middle of an FTP download, of course. But then, nobody (so far) seems to have thought of it that even Win9x and NT can let the harddisk spin down when it hasn't been accessed for a couple of minutes (and that's also including the disk containing the paging file...), so it must be easy to survive ;-) -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <366b3255.941270842@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 60 Date: Fri, 04 Dec 1998 19:08:42 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 912798577 129.85.24.56 (Fri, 04 Dec 1998 14:09:37 EDT) NNTP-Posting-Date: Fri, 04 Dec 1998 14:09:37 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!news.maxwell.syr.edu!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >alt.folklore.computers: > >> In article <36677771.23568198@news.vip.net>, >> genew@vip.net (Gene Wirchenko) wrote: >> > This is impressive. Do you have any details on how it was done? >> >The post before said that VMS managed even with the VM/paging disk out >> >of action. How? >> >> Sigh! It's called timesharing. That's an aspect of implementation >> of timesharing. The software needs to be robust, no matter what >> those pesky users do. Hardware off-line interrupts are a piece of >> cake in comparison. TOPS10 knew how to do all that in the 70s (if >> not the 60s). > >Sorry if this is a stupid remark, but I don't see in what way the >fact that it's a timesharing system can have anything to do with >being able to continue whatever it's doing when the system disk >goes off-line (or after it comes back on)? _Proper implementation_ of timesharing >For all I know, timesharing is just another way to implement >multitasking (in fact, resembling what we call preemptive >multitasking today, in that it slices up available CPU time over >different users or processes). > >I guess I'm about 5 years too young to really know timesharing >though. > >BTW, some of my first steps were made on RSX-11 - and I seem to >remember (vaguely, after 20 years) entering the building one >morning, finding both harddisks (thus including the system disk) >powered off: after switching them on, the system ran normally. > >And I wouldn't even find it very strange: AFAIK, the system did >nothing in the background like today's systems do (in a way - >sorry for the comparison - it was more like DOS rather than >Windows in that aspect, at least that's how it appeared to be on >that particular system: I never saw any disk activity when I >wasn't running anything). Again, this particular system may have an empty idle process, but RSX-11 per se was quite capable of running detached batch jobs. >Not quite the same as being able to survive power cycling in the >middle of an FTP download, of course. > >But then, nobody (so far) seems to have thought of it that even >Win9x and NT can let the harddisk spin down when it hasn't been >accessed for a couple of minutes (and that's also including the >disk containing the paging file...), so it must be easy to >survive ;-) Smiley noted. Comparison of Windows with real operating system excused. [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### From: ftit@engin.umich.edu (Sergej Roytman) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 19:13:28 GMT Organization: University of Michigan, College of Engineering Lines: 32 Message-ID: <749c8o$io4@srvr1.engin.umich.edu> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> NNTP-Posting-Host: az.engin.umich.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsxfer.itd.umich.edu!srvr1.engin.umich.edu!ftit In article <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de>, >"Strange, but consistent" is IMO typical for a Unix person's experiences >with VMS. "Valid, but verbose" was mine. Of course, I'm a guy who likes to call his compiled programs "p", because it's easy to place an index finger on the P-key and a pinky on the return key... > Of course, VMS beginners from other backgrounds may find >commands like "dir" (or "directory") to list a directory or "copy" to >copy files not *that* strange, perhaps not even "run" to run executable >programs (necessary if they are not installed as commands). My response was: what's the connection with MessyDOS? Hm, since this is the folklore newsgroup, what _is_ the connection? At a guess, CP/M->QDOS->The Program-Loader of the Beast? >In lots of ways VMS is very consistent, and in lots of ways it is very >flexible. These are, though, very different from the consistent (where >existent) and flexible ways of Unix. This is the reason that Unix people >very often find VMS strange and develop an intense dislike towards VMS. I did not dislike VMS in the brief time that I worked with it. Yes, I thought it strange, but it looked like something humans could come up with if they hadn't come up with Unix. IBM's operating systems, on the other hand, all seem to have been written by Martians! -- Sergej Roytman ###### From: ftit@engin.umich.edu (Sergej Roytman) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 19:21:39 GMT Organization: University of Michigan, College of Engineering Lines: 24 Message-ID: <749co3$j9h@srvr1.engin.umich.edu> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> NNTP-Posting-Host: az.engin.umich.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsxfer.itd.umich.edu!srvr1.engin.umich.edu!ftit In article <366b3255.941270842@Rockyd>, Alexandre Pechtchanski wrote: >On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: >>Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >>alt.folklore.computers: >>> In article <36677771.23568198@news.vip.net>, >>> genew@vip.net (Gene Wirchenko) wrote: >>> > This is impressive. Do you have any details on how it was done? >>> >The post before said that VMS managed even with the VM/paging disk out >>> >of action. How? >>> Sigh! It's called timesharing. That's an aspect of implementation >>Sorry if this is a stupid remark, but I don't see in what way the >>fact that it's a timesharing system can have anything to do with >>being able to continue whatever it's doing when the system disk >>goes off-line (or after it comes back on)? >_Proper implementation_ of timesharing For those yunguns among us who think Unix is pretty cool (well, it is!), please clarify. I mean, I can see things working OK if everything that's needed from the disk is in cache, but what happens the first time the OS decides it's time to page something out? -- Sergej Roytman ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <36683c0a.1813126@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 44 Date: Fri, 04 Dec 1998 19:56:23 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 912801376 129.85.24.56 (Fri, 04 Dec 1998 14:56:16 EDT) NNTP-Posting-Date: Fri, 04 Dec 1998 14:56:16 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!cyclone.news.idirect.com!island.idirect.com!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On 4 Dec 1998 19:21:39 GMT, ftit@engin.umich.edu (Sergej Roytman) wrote: >In article <366b3255.941270842@Rockyd>, >Alexandre Pechtchanski wrote: >>On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: >>>Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >>>alt.folklore.computers: >>>> In article <36677771.23568198@news.vip.net>, >>>> genew@vip.net (Gene Wirchenko) wrote: >>>> > This is impressive. Do you have any details on how it was done? >>>> >The post before said that VMS managed even with the VM/paging disk out >>>> >of action. How? >>>> Sigh! It's called timesharing. That's an aspect of implementation >>>Sorry if this is a stupid remark, but I don't see in what way the >>>fact that it's a timesharing system can have anything to do with >>>being able to continue whatever it's doing when the system disk >>>goes off-line (or after it comes back on)? >>_Proper implementation_ of timesharing > >For those yunguns among us who think Unix is pretty cool (well, it >is!), please clarify. I mean, I can see things working OK if >everything that's needed from the disk is in cache, but what happens >the first time the OS decides it's time to page something out? I can't say anything about TOPS (being a yungun myself in this respect ;-) but my understanding of VMS internal workings is that it won't. The event sequence is approximately like this: 1. Program asks for more data/overlay a) program posts a request to operating system to do that b) program provides a flag for notification c) system queues the request and puts program to sleep 2. System determines that there is a request in the queue that needs servicing. System checks resource availability to service the request. Check returns "resources unavailable". System goes to the next request in queue. This is necessarily oversimplified, but I hope conveys the idea. The programs that don't need unavailable resources continue to run. The programs that _do_ need these resources sleep waiting for their availability. (Resource availability is determined by similar mechanisms, so the system won't try to access hard drive to see that it is available for each request - no busy I/O loops, no sirree). [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### Newsgroups: alt.folklore.computers From: eric@fudge.uchicago.edu (Eric Fischer) Subject: Re: Is VMS taboo, or what? X-Nntp-Posting-Host: fudge.uchicago.edu Message-ID: Sender: news@midway.uchicago.edu (News Administrator) X-Newsposter: Pnews 4.0-test51 (15 Jan 97) Organization: The University of Chicago References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> Date: Fri, 4 Dec 1998 20:09:53 GMT Lines: 14 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!uchinews!not-for-mail Sergej Roytman wrote: > My response was: what's the connection with MessyDOS? Hm, since this > is the folklore newsgroup, what _is_ the connection? At a guess, > > CP/M->QDOS->The Program-Loader of the Beast? The part of the connection left out is: +----> CP/M ---> MS-DOS earlier DEC operating systems ---| +----> VMS eric ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 21:46:44 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 21 Message-ID: <749l84$7vd$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <749co3$j9h@srvr1.engin.umich.edu>, Sergej Roytman wrote: >For those yunguns among us who think Unix is pretty cool (well, it >is!), please clarify. I mean, I can see things working OK if >everything that's needed from the disk is in cache, but what happens >the first time the OS decides it's time to page something out? Speaking as someone who worked for many year on diskless UNIX workstations, the answer is easy, and obvious. The only reason to page something out is that a process wants to page something in. The OS just suspends the process requesting that page until the disk (or in the case of a diskless workstation, the page file server) is back in working order. Programs that aren't requesting "paged-out" pages should continue to run happily. I can't think of any reason for a real OS to crash just because it can't reach the page disk. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 4 Dec 1998 21:52:25 GMT Organization: University of Illinois at Urbana-Champaign Lines: 23 Message-ID: <749lip$gjf$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk>, Edward John M. Brocklesby wrote: >On about Thu, 3 Dec 1998 00:02:00 +1030, The XO might have written: >[snip story about HD] > >I have a Linux box here that has a single 8.5Gb EIDE HD running the whole >thing. Occasionally, I do something stu^H^H^Hinteresting, like drop it, >or turn it off accidently (it's on a seperate PSU to the computer), and >whatever I do, Linux manages to recover from it fine. (Unlike NetBSD, >which just hangs when I turn the HD off). > >That's on Amiga hardware, don't know if it would work on PC's. I tried doing this to my Linux machine here (PC), I pulled out the hard drive cable, did some file I/O, replugged it in, and I saw one of the most spectacular Linux crashes ever. All of the windows (displaying with X remotely to my Windows workstation) starting closing one by one, my telnet sessions to the machine started closing, and the console completely locked up. Of course, nobody expected Linux to be as robust as VMS, did they? ;-) -- Terry ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 00:30:32 GMT Organization: Posted via Supernews, http://www.supernews.com - Discussions start here! Lines: 60 Message-ID: <366858fd.4481731@news.vip.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.27 X-Trace: 912817678 A01OARAUVD41BCCD1C usenet78.supernews.com X-Complaints-To: newsabuse@supernews.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!news-feed.inet.tele.dk!bofh.vszbr.cz!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >On 4 Dec 1998 19:21:39 GMT, ftit@engin.umich.edu (Sergej Roytman) wrote: > >>In article <366b3255.941270842@Rockyd>, >>Alexandre Pechtchanski wrote: >>>On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: >>>>Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >>>>alt.folklore.computers: >>>>> In article <36677771.23568198@news.vip.net>, >>>>> genew@vip.net (Gene Wirchenko) wrote: >>>>> > This is impressive. Do you have any details on how it was done? >>>>> >The post before said that VMS managed even with the VM/paging disk out >>>>> >of action. How? >>>>> Sigh! It's called timesharing. That's an aspect of implementation >>>>Sorry if this is a stupid remark, but I don't see in what way the >>>>fact that it's a timesharing system can have anything to do with >>>>being able to continue whatever it's doing when the system disk >>>>goes off-line (or after it comes back on)? >>>_Proper implementation_ of timesharing >> >>For those yunguns among us who think Unix is pretty cool (well, it >>is!), please clarify. I mean, I can see things working OK if >>everything that's needed from the disk is in cache, but what happens >>the first time the OS decides it's time to page something out? > >I can't say anything about TOPS (being a yungun myself in this respect ;-) but >my understanding of VMS internal workings is that it won't. The event sequence >is approximately like this: >1. Program asks for more data/overlay > a) program posts a request to operating system to do that > b) program provides a flag for notification > c) system queues the request and puts program to sleep >2. System determines that there is a request in the queue that needs servicing. >System checks resource availability to service the request. Check returns >"resources unavailable". System goes to the next request in queue. > >This is necessarily oversimplified, but I hope conveys the idea. The programs >that don't need unavailable resources continue to run. The programs that _do_ >need these resources sleep waiting for their availability. (Resource >availability is determined by similar mechanisms, so the system won't try to >access hard drive to see that it is available for each request - no busy I/O >loops, no sirree). I thought that was how it would be. What I was really worried about was how it would handle when a swap was needed, especially a swap IN. So, in the example given, the user was lucky that no swapping was required? I understand that you could probably write the system so swaps were less likely to be needed. Was this done? Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: gerglery@usa.net (Fluffy) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 01:21:40 GMT Organization: If one may meow, all may meow. Message-ID: <74a1r4$e5s$1@meowhost.meow.invalid> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> X-Trace: mail2news.demon.co.uk 912828919 mail2news:1373 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!out2.ibm.net!meowhost.meow.invalid X-P-Meow: Meow Mail-Copies-To: never X-URL: http://members.tripod.com/~gerglery/ X-Newsreader: slrn (0.9.5.3-canlock UNIX) Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Gene Wirchenko wrote: > I thought that was how it would be. What I was really worried > about was how it would handle when a swap was needed, especially a > swap IN. So, in the example given, the user was lucky that no > swapping was required? It wasn't luck. It was designed to degrade gracefully. > I understand that you could probably write the system so swaps > were less likely to be needed. Was this done? Yup. VMS' memory model is a lot more sophisticated than simply swapped out or in core. It's broken down into a hierarchy of 'pools' in which the critical stuff is pretty much always directly addressable. Figuring out optimal sizes for those chunks of memory and the counters that contol them is one of the more, um, enjoyable parts of making that OS run well. There's an autogen utility that can calculate some pretty good guesses, but sometimes you just need to get in there and figure it out for yourself :/ ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 01:44:11 GMT Message-ID: <74a35b$hrf$8@irk.zetnet.co.uk> References: <748ist$lfi$1@strato.ultra.net> NNTP-Posting-Host: man-059.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 912822251 18287 194.247.41.73 (5 Dec 1998 01:44:11 GMT) NNTP-Posting-Date: 5 Dec 1998 01:44:11 GMT X-Everything: Net-Tamer V 1.08X Lines: 18 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-04 jmfbahciv@aol.com said: :Really? My interpretation of those disclaimers was WYSIWYG, :promising nothing about performance. Yes, but those disclaimers aren't valid under UK case law. It's been established that if you're selling something, (a) it *must* be able to do the job it's being sold for, and (b) the buyer must understand what they're getting. In fact, Microsoft were clobbered for some aspect of their trading practices a while back (not necessarily this one, but). I would return a copy of NT myself as of unmerchantable quality (as per Robert's suggestion) but since I won't alllow it in the house... (Yay! I now run Microsoft-free - and it's still DOS!) -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 01:44:14 GMT Lines: 36 Message-ID: <74a35e$hrf$9@irk.zetnet.co.uk> References: <748o17$v01$1@strato.ultra.net> NNTP-Posting-Host: man-059.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 912822254 18287 194.247.41.73 (5 Dec 1998 01:44:14 GMT) NNTP-Posting-Date: 5 Dec 1998 01:44:14 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-04 jmfbahciv@aol.com said: :> OTOH if you describe something as an operating system, then it's :>jolly well got to be one, or you can be prosecuted under the TDA. :It only has to function as an OS. It still says nothing about :performance quality, e.g. not crashing a lot. But a case could pretty easily be made for the instability of Win95, and the consequent loss of functionality and performance, rendered the product unsuitable for its purported function. (It would have been easier with Win3.1, perhaps, but now there is no alternative.) :> Similarly, :>if inside the shrink wrap it says, "you aren't allowed to make :>backups," then it's too late, you can make as many backups as you :like. > :Backups of what? The system disk? Hmmm...there's really bad :implications there. Indeed. Fortunately, if it's inside the box, it doesn't count - you have to be able to examine the conditions of sale before purchase (and that doesn't mean another copy from another box, I guess). Another example is MS' practice of putting the licence agreement on the CD, with the installation package. (In fact, the concept of a "sealed box" might itself be open to question.) The real problem with all of this is that with most software, you're actually only buying a licence to corrupt your hard drive with it. That potentially limits your rights to return or replacement of media only, and then only if faulty. That's as good a reason as any to only use Open Source software. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 05:11:41 GMT Organization: Micro Focus Inc. Lines: 41 Message-ID: <74afad$nk8@news1.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <749l84$7vd$1@agate.berkeley.edu> NNTP-Posting-Host: p-704.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!128.223.220.30!logbridge.uoregon.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!mww In article <749l84$7vd$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) writes: > > Speaking as someone who worked for many year on diskless UNIX workstations, > the answer is easy, and obvious. The only reason to page something out > is that a process wants to page something in. The OS just suspends the > process requesting that page until the disk (or in the case of a diskless > workstation, the page file server) is back in working order. Programs that > aren't requesting "paged-out" pages should continue to run happily. I can't > think of any reason for a real OS to crash just because it can't reach the > page disk. Unless it has a pageable kernel, like AIX versions 3 and up, and the disk fails when some crucial portion of the kernel is paged out. (Though of course that's a fatal page in failure, not a problem with paging out.) Some may not consider AIX "a real OS" - or, more seriously, that the potential exposure of a pageable kernel isn't worth the advantages (like more efficient physical memory usage and simplified design of some kernel components). IMO, considering the average reliability of disks these days, the low disk prices that make redundant disk arrays widely available, and the availability of enhanced-reliability options (like HACMP, in the AIX case) it's a reasonable trade-off. ObVMS: I'm a Unix type myself, but in college I did a fair bit of assembly programming on VAX VMS (Northeastern's CS program had two required assembly courses, taught with VMS; I bet there isn't even one assembly course required for CS majors there now). It's a great platform for assembly - all those CISC microprogrammed opcodes, a decent macro language, and some really useful macros in the standard library (gotta love that $UNWIND). Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University He described a situation where a man is there to feed a dog and the dog is there to keep the man from touching the equipment. -- Anthony F. Giombetti ###### Reply-To: "Dennis J. Minette" From: "Dennis J. Minette" References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> Subject: Re: Is VMS taboo, or what? Date: Sat, 5 Dec 1998 05:58:36 -0500 Lines: 16 Organization: Minette Data Systems, Inc. X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Message-ID: Newsgroups: alt.folklore.computers NNTP-Posting-Host: 1Cust217.tnt1.sarasota.fl.da.uu.net [153.37.162.217] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!207.68.152.14!upnetnews04!upnetnews03 Alexandre Pechtchanski wrote in message <366b3255.941270842@Rockyd>... >On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: < BIG snippage here > >Smiley noted. Comparison of Windows with real operating system excused. > Did you mean reel operating system like TOS (as in Tape Operating System)? ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 10:19:29 -0600 Organization: MindSpring Enterprises Lines: 14 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> NNTP-Posting-Host: d1.56.d2.91 X-Server-Date: 5 Dec 1998 16:19:27 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <366858fd.4481731@news.vip.net>, genew@vip.net wrote: > I thought that was how it would be. What I was really worried >about was how it would handle when a swap was needed, especially a >swap IN. So, in the example given, the user was lucky that no >swapping was required? Actually, he answered your question. A "swap in" request is a request for a resource. GO TO the part of his note which spoke of what happens when a requested resource is unavailable. Yes, I know I've created an unstructured message. Life is tough. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 10:23:41 -0600 Organization: MindSpring Enterprises Lines: 10 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <749l84$7vd$1@agate.berkeley.edu> <74afad$nk8@news1.newsguy.com> NNTP-Posting-Host: d1.56.d2.91 X-Server-Date: 5 Dec 1998 16:23:40 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74afad$nk8@news1.newsguy.com>, mww@microfocus.com (Michael Wojcik) wrote: >ObVMS: I'm a Unix type myself, but in college I did a fair bit of >assembly programming on VAX VMS (Northeastern's CS program had two >required assembly courses, taught with VMS; When-abouts were you there? I was there in 1982-1983. -- Howard S Shubs The Denim Adept ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 10:49:08 GMT Organization: . Lines: 27 Message-ID: <36690870.1413602@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> NNTP-Posting-Host: pool02b-194-7-146-21.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news.gsl.net!gip.net!newsfeed.tli.de!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach ftit@engin.umich.edu (Sergej Roytman) on 4 Dec 1998 19:13:28 GMT to alt.folklore.computers: > My response was: what's the connection with MessyDOS? Hm, since this > is the folklore newsgroup, what _is_ the connection? At a guess, > > CP/M->QDOS->The Program-Loader of the Beast? Not: it's the program loader of win9x, but win9x doesn't have any VMS roots. NT does, but it doesn't use DOS as a loader (the boot sector may be similar, but that's about it). In the worst case, you can expand Eric's map to: +--> CP/M --> DOS --> (Win3) -+--> Win9X earlier DEC OSes --| V +--> VMS -------------------->+--> NT Calling Win3 an OS may be a bit too much, but I put it there because NT inherited its API. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 10:52:01 +0000 Organization: Tanglewood Message-ID: <36691051.D3041D1C@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912856897 nnrp-02:18838 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 18 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Sergej Roytman wrote: > My response was: what's the connection with MessyDOS? Hm, since this > is the folklore newsgroup, what _is_ the connection? At a guess, > > CP/M->QDOS->The Program-Loader of the Beast? PDP/11-DOS -> CP/M->QDOS->The Program-Loader of the Beast A *lot* of people have forgotten that the PDP/11 ran DOS (the real one, not M$) for several years before RSX was invented. A bit like Windows and Linux really... -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 11:06:57 +0000 Organization: Tanglewood Message-ID: <366913D1.986841E5@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912856898 nnrp-02:18838 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 26 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Gene Wirchenko wrote: > I understand that you could probably write the system so swaps > were less likely to be needed. Was this done? Yes, and how it does it is very clever. Basically (scratches head, tries to remember DM from BTDOT) VMS knows when a page has been written to by a process, and so knows that some pages don't match what is on disk, these pages have to be rewritten if they are to go out of memory. What it then does is remove pages from process working sets, but leave them in memory. If the process asks for them again, it simply reattaches them to the working set, if the process doesn't ask for them back, then in time they are treated as free memory and can be allocated to other processes. If the original process asks for the page again, then new memory is allocated, and the page read in again from disk. What this means in practice is that pages which aren't being used (such as initialisation code) get nibbled out of the working set, leaving only the active pages in memory. Or something like that, it's been a *long* time... -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 11:12:57 GMT Organization: Daedalus Consulting Lines: 21 Message-ID: <74b4fp$54m$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <365d9b9c.101373846@news.mcs <36678551.2510824@news.mcs.net> <749225$1kv$1@news.han.telia.se> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Peter Ljungberg wrote: >I once by mistake, pulled the SDI-cable (black cable between >computer and/or disk/controller) to the systemdisk, heard >beeping from the console, which said that DUA0: (the systemdisk) >was offline, so I placed it back where it was before and everything >was as before.... I walked into the machine room a few times to be greeted by 20-odd RA81 & RA82 fault lights (they're a particularly nasty shade of red), including the system disks, after power flicks. They were generally OK, just freaked. Spin 'em back up, and the (VMS) VAXes came back to life as if nothing had happened. On the other hand, I once accidentally powered down a couple of RA90s attached to an Ultrix box (the drives were on the same power controller as a defunct machine in an adjacent cabinent that I was cannibalising). Boy, was I popular. Ultrix fell flat and hard on its face, requiring a a two-hour manual fsck to get it back up. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jmfbah@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 98 12:12:56 GMT Organization: UltraNet Communications, Inc. Lines: 52 Message-ID: <74b9ad$anb$1@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> NNTP-Posting-Host: d17.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 5 Dec 1998 12:35:25 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!feed1.news.rcn.net!rcn!news.ultranet.com!d17 In article <36691be2.2492303@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >alt.folklore.computers: > >> In article <36677771.23568198@news.vip.net>, >> genew@vip.net (Gene Wirchenko) wrote: >> > This is impressive. Do you have any details on how it was done? >> >The post before said that VMS managed even with the VM/paging disk out >> >of action. How? >> >> Sigh! It's called timesharing. That's an aspect of implementation >> of timesharing. The software needs to be robust, no matter what >> those pesky users do. Hardware off-line interrupts are a piece of >> cake in comparison. TOPS10 knew how to do all that in the 70s (if >> not the 60s). > >Sorry if this is a stupid remark, but I don't see in what way the >fact that it's a timesharing system can have anything to do with >being able to continue whatever it's doing when the system disk >goes off-line (or after it comes back on)? It's not a stupid remark...see below (you're confused :-)). > >For all I know, timesharing is just another way to implement >multitasking (in fact, resembling what we call preemptive >multitasking today, in that it slices up available CPU time over >different users or processes). Multitasking is _not_ timesharing in our definition. If you had added the words "making all resources of the computer system available to each user during its runtime slice" in your parentheses, one would have timesharing. Note that when I say "all resources", I also include all memory which is one of the main differences between timesharing and multitasking. > >I guess I'm about 5 years too young to really know timesharing >though. > >BTW, some of my first steps were made on RSX-11 - and I seem to >remember (vaguely, after 20 years) entering the building one >morning, finding both harddisks (thus including the system disk) >powered off: after switching them on, the system ran normally. Which RSX did you use? /BAH Valid e-mail address since either the spammers have forgotten about me or AOL's mailing system didn't work yesterday. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 98 12:33:58 GMT Organization: UltraNet Communications, Inc. Lines: 100 Message-ID: <74bahs$anb$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> NNTP-Posting-Host: d17.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 5 Dec 1998 12:56:28 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d17 In article <36683c0a.1813126@Rockyd>, alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >On 4 Dec 1998 19:21:39 GMT, ftit@engin.umich.edu (Sergej Roytman) wrote: > >>In article <366b3255.941270842@Rockyd>, >>Alexandre Pechtchanski wrote: >>>On Fri, 04 Dec 1998 18:19:48 GMT, lucvdv@null.net (Luc Van der Veken) wrote: >>>>Also sprach jmfbahciv@aol.com on Fri, 04 Dec 98 11:42:08 GMT to >>>>alt.folklore.computers: >>>>> In article <36677771.23568198@news.vip.net>, >>>>> genew@vip.net (Gene Wirchenko) wrote: >>>>> > This is impressive. Do you have any details on how it was done? >>>>> >The post before said that VMS managed even with the VM/paging disk out >>>>> >of action. How? >>>>> Sigh! It's called timesharing. That's an aspect of implementation >>>>Sorry if this is a stupid remark, but I don't see in what way the >>>>fact that it's a timesharing system can have anything to do with >>>>being able to continue whatever it's doing when the system disk >>>>goes off-line (or after it comes back on)? >>>_Proper implementation_ of timesharing >> >>For those yunguns among us who think Unix is pretty cool (well, it >>is!), please clarify. I mean, I can see things working OK if >>everything that's needed from the disk is in cache, but what happens >>the first time the OS decides it's time to page something out? > >I can't say anything about TOPS (being a yungun myself in >this respect ;-) but my understanding of VMS internal workings >is that it won't. The event sequence is approximately like this: >1. Program asks for more data/overlay > a) program posts a request to operating system to do that > b) program provides a flag for notification > c) system queues the request and puts program to sleep >2. System determines that there is a request in the queue >that needs servicing. System checks resource availability to >service the request. Check returns >"resources unavailable". System goes to the next request in queue. Or, if the program requested it, returns control to the program to do other stuff; we called it non-blocking I/O (which was real dangerous to do), buffered mode I/O which allowed the program to continue doing it I/O thing without having to wait for the device to complete a particular I/O UUO, and IPCF which allowed communication between processes without having to wait for an answer before doing other work. > >This is necessarily oversimplified, but I hope conveys the >idea. You're doing just fine :-). >The programs that don't need unavailable resources >continue to run. The programs that _do_ need these resources >sleep waiting for their availability. (Resource availability >is determined by similar mechanisms, so the system won't try to >access hard drive to see that it is available for each >request - no busy I/O loops, no sirree). Then, an OS must consider errors, hardware, software, stupidity, etc. Take the simple case, a disk goes off-line. The hardware flags an off-line interrupt. If it's an unexpected off-line, the monitor (that's what we called the part of the OS that was always resident in core) notices because that's its job. It does all the appropriate things, like kick the operator in the butt (we did it with , , ) to fix the problem, gracefully queue up the jobs who need to wait until the hardware is fixed, and rudely stop the jobs whose function can't be recovered or continued because of the hardware fault. Now, the differences between operating systems was the philosophy. One of the things that I don't like about UNIX is that part of the philosophy that says, "We'll take care of the writing of the data in our own sweet time. You (the program) can just assume that it's written out." In TOPS10, if the program said, "I want it written out and let me know when it's cast in iron.", the monitor obliged. In all cases, all I/O was written when the program called it quits. Now, when one begins to think of terminal I/O, the fit can hit the shan. Our implementations of networking also encountered similar difficulties because now one has to figure out whose allowed to have the sayso on that other remote system (which may not have users (I'm talking about terminal concentrators essentially). Now there are reasons for the differences in philosophies which can produce great, interesting, and heated discussions :-). We ususally had such discussions when designing new software concepts. It could be bloody :-). /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 98 12:36:00 GMT Organization: UltraNet Communications, Inc. Lines: 23 Message-ID: <74balk$anb$3@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36691051.D3041D1C@tnglwood.demon.co.uk> NNTP-Posting-Host: d17.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 5 Dec 1998 12:58:28 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!feeder.qis.net!newsfeed.xcom.net!news.ultranet.com!d17 In article <36691051.D3041D1C@tnglwood.demon.co.uk>, Robert Billing wrote: >Sergej Roytman wrote: > >> My response was: what's the connection with MessyDOS? Hm, since this >> is the folklore newsgroup, what _is_ the connection? At a guess, >> >> CP/M->QDOS->The Program-Loader of the Beast? > > PDP/11-DOS -> CP/M->QDOS->The Program-Loader of the Beast > > A *lot* of people have forgotten that the PDP/11 ran DOS (the real one, >not M$) for several years before RSX was invented. A bit like Windows >and Linux really... > I sure am one of them. I encountered DOS in the late 70s but used RSXes (many flavors of it) in the early 70s. When was DOS developed? /BAH Subtract a hundred and four for e-mail. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 14:21:11 +0000 Organization: Tanglewood Message-ID: <36694157.59552017@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36691051.D3041D1C@tnglwood.demon.co.uk> <74balk$anb$3@ligarius.ultra.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912877312 nnrp-01:16824 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 19 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!bullseye.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail jmfbahciv@aol.com wrote: > I sure am one of them. I encountered DOS in the late 70s but used > RSXes (many flavors of it) in the early 70s. When was DOS developed? Well, it was relatively new and at V8.2 when I used it in my first paid programming job on a PDP/11-40 in 1973. I would suspect that the original release was around 1971. Major releases after this were 8.8, 9.0 and 10.x which was then replaced by 10. It supported asynchronous I/O, and various multi-tasking kernels were built on top of it for ad-hoc systems, until it was finally replaced by RSX. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 14:24:11 +0000 Organization: Tanglewood Message-ID: <3669420B.F36042F4@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <74bahs$anb$2@ligarius.ultra.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 912877313 nnrp-01:16824 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 14 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail jmfbahciv@aol.com wrote: > Or, if the program requested it, returns control to the program to > do other stuff; we called it non-blocking I/O (which was real > dangerous to do), buffered mode I/O which allowed the program Whaddyamean dangerous? Real Programmers always do non-blocking I/O, it lets you get more done that way. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 17:12:56 GMT Organization: . Lines: 41 Message-ID: <366a6629.4860679@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <74b9ad$anb$1@ligarius.ultra.net> NNTP-Posting-Host: pool02b-194-7-145-235.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!rill.news.pipex.net!pipex!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach jmfbah@aol.com on Sat, 05 Dec 98 12:12:56 GMT to alt.folklore.computers: > Note that when > I say "all resources", I also include all memory which is one > of the main differences between timesharing and multitasking. It's actually the only difference I knew of - that and I think I read somewhere that the time slices were bigger. > Which RSX did you use? > Sorry, but here comes another stupid question: Was there more than one? (slightly embarrassed emoticon here) In other words, I wouldn't know if you asked me with a stick. I was a student working on my [how do you call that in English again - over here we say thesis] in those days, and all I really knew about RSX was how to log on, edit a source, fire up the Fortran compiler and run something called TKB - the 'task builder', which I suspected to be something like a linker. And run the compiled program. The PDP11 (-45 IIRC) was used as a digital assistant to an analog computer (Applied Dynamics Five), and that's where I did most of my work on. [That's what you get when you work in a government-controlled research facility: you can use a PDP-11 - two rack cabinets big, each of which half empty - for no other purpose than to store and set the potentiometer settings of an analog computer, so you don't have to change them by hand each time you load another model :-) Sounds like a waste, but when you consider what the AD/5 cost compared to a PDP-11, it almost became 'an affordable extra'] -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: huge@nospam.demon.co.uk (Hugh Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 5 Dec 1998 18:37:24 GMT Organization: Piglet's Pickles and Preserves Message-ID: <74buh4$sd@axalotl.demon.co.uk> References: <366a6629.4860679@news.uunet.be> Reply-To: huge@nospam.demon.co.uk NNTP-Posting-Host: axalotl.demon.co.uk X-NNTP-Posting-Host: axalotl.demon.co.uk:158.152.24.143 X-Trace: news.demon.co.uk 912888078 nnrp-10:22674 NO-IDENT axalotl.demon.co.uk:158.152.24.143 X-Complaints-To: abuse@demon.net x-no-archive: yes Lines: 18 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!axalotl.demon.co.uk!axalotl!usenet In article <366a6629.4860679@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) writes: >> Which RSX did you use? >> > >Sorry, but here comes another stupid question: >Was there more than one? (slightly embarrassed emoticon here) He-he-he. -S, -M, -M+ & -D, that I know of. -- "The road to Paradise is through Intercourse." The uk.transport FAQ; http://www.axalotl.demon.co.uk/transport/FAQ.html [Substitute "axalotl" for "nospam" to email me] ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 05 Dec 1998 19:32:46 -0600 Organization: MindSpring Enterprises Lines: 21 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <749co3$j9h@srvr1.engin.umich.edu> <749l84$7vd$1@agate.berkeley.edu> <74afad$nk8@news1.newsguy.com> <74cit5$n6g$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: d1.56.d0.74 X-Server-Date: 6 Dec 1998 01:32:45 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <74cit5$n6g$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz (Don Stokes) wrote: >Agreed, the VAX under VMS is a lovely machine to write assembly language >on, the only real black spot being the byte-relative conditional >branches. There are a couple of traps -- the character string >instructions using the low-order registers as scratchpad space is a >nuisance and offers a nasty surprise for the novice, although the >surprise would be mitigated somewhat if the architecture book pointed >this out in the specific descriptions of the instructions rather than >burying the info in the general description of the instruction set. It's right there on page 283, under the description of the MOVC instruction. " 291, " " " " " MOVTC " etc... You -do- have the "white" version of the VAX Architecture Handbook, right? The other one should be called the VAX Arghitecture Handbook; it's kinda sorta the same thing but... -- Howard S Shubs The Denim Adept ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 6 Dec 1998 00:25:09 GMT Organization: Daedalus Consulting Lines: 58 Message-ID: <74cit5$n6g$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <749co3$j9h@srvr1.engin.umich.edu> <749l84$7vd$1@agate.berkeley.edu> <74afad$nk8@news1.newsguy.com> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74afad$nk8@news1.newsguy.com>, Michael Wojcik wrote: >Unless it has a pageable kernel, like AIX versions 3 and up, and the >disk fails when some crucial portion of the kernel is paged out. >(Though of course that's a fatal page in failure, not a problem with >paging out.) The VMS kernel has always had pageable parts -- anything that has to run or be accessed at elevated IPL has to be non-pageable (specifically from IPL > 2, which permits things like special kernel ASTs to run in pageable space while being protected from accidents by elevated IPL), but merely running in kernel mode isn't any excuse for locking memory down for your exclusive use. There's a difference between getting a data error where the hardware reports that the disk is there but the data is gone, and the hardware reporting that the device has gone offline or just not responding at all. A permanent disk error while paging in a piece of kernel code will kill VMS for sure, and errors at lower access levels will cause various levels of damage depending on where they happened, typically the death of the running process or image, or at least the signalling of an error to any active signal handlers. But a disk request not being answered at all, or the controller saying the disk has gone offline, can simply be handled by not completing the request until the disk comes back -- since each device has its own queue of requests, this works by just forgetting about the thing completely until the driver gets an interrupt from the device indicating completion of the request or that it's back and ready for more requests. Disks going offline and being signalled as such (MSCP hardware is generally pretty good at this) usually result in the device being placed in "mount verification". A disk in mount verification just doesn't get handed requests. Mount verification times out after a while (the timeout is a SYSGEN parameter, dynamic too so it can be changed on the fly) after which the driver gives up on the disk and unmounts it. Usually it's set to the default of one hour. Again, once the disk is signalled as being online, mount verification is completed and requests can be queued again. >ObVMS: I'm a Unix type myself, but in college I did a fair bit of >assembly programming on VAX VMS (Northeastern's CS program had two >required assembly courses, taught with VMS; I bet there isn't even >one assembly course required for CS majors there now). It's a great >platform for assembly - all those CISC microprogrammed opcodes, a >decent macro language, and some really useful macros in the standard >library (gotta love that $UNWIND). Agreed, the VAX under VMS is a lovely machine to write assembly language on, the only real black spot being the byte-relative conditional branches. There are a couple of traps -- the character string instructions using the low-order registers as scratchpad space is a nuisance and offers a nasty surprise for the novice, although the surprise would be mitigated somewhat if the architecture book pointed this out in the specific descriptions of the instructions rather than burying the info in the general description of the instruction set. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 6 Dec 1998 01:26:08 +0100 Organization: [Posted via] Interactive Networx Lines: 24 Message-ID: <1djko2i.1mnojj0nf1bccN@n245-102.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> NNTP-Posting-Host: n245-102.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.pop.de!wuff.mayn.de!unlisys!news.snafu.de!jnickelsen Sergej Roytman wrote: > IBM's operating systems, on the other hand, all seem to have been > written by Martians! This has at least been told of AIX. Said Martians were given the specification of Unix and were told to come up with something alike. As we all know by now, the specification was not quite complete. Things like this do happen, though. I remember reading, I think in Guy L. Steele's article on the development of Lisp for the second History of Programming Languages Conference, that after the first Common Lisp specification had been published, a group from Kyoto wrote Kyoto Common Lisp (KCL). This implementation behaved quite differently from what had been common practice in the Lisp community, due to the fact that the KCL implementors had used the specification as the source without having had much contact to the Lisp community and its unwritten (apart from source code) traditions. According to GLS this pointed to the holes in the CL specification quite neatly. -- Juergen Nickelsen ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 6 Dec 1998 09:02:45 GMT Organization: Daedalus Consulting Lines: 34 Message-ID: <74dh7l$h0t$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74afad$nk8@news1.newsguy.com> <74cit5$n6g$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Howard S Shubs wrote: >>surprise would be mitigated somewhat if the architecture book pointed >>this out in the specific descriptions of the instructions rather than >>burying the info in the general description of the instruction set. > >It's right there on page 283, under the description of the MOVC instruction. > " 291, " " " " " MOVTC " > etc... > >You -do- have the "white" version of the VAX Architecture Handbook, right? >The other one should be called the VAX Arghitecture Handbook; it's kinda >sorta the same thing but... Yeah, I was thinking of the 1986 rice-paper edition, which does have (not terribly useful) stuff on subsetting and so-on not in the earlier books. You're right, the earlier books do describe things a bit better. 'course my '77 and '79 books omit little things like G- and H-floating datatypes as well. My '82 "VAX-11 Architeture Reference Manual", in all its 10-pt Courier glory does include these, and goes into excruciating detail. (BTW: the page numbers are a bit version-specific; the 1977 "VAX 11/780 ARCHITECTURE HANDBOOK VOL. 1" book has MOVC on P. 9-2 and MOVTC on 9-4. I've never seen a VOL. 2 of this book, but the '79 book omits "/780" from the title, and comes with the "VAX 11/780 HARDWARE HANDBOOK" that describes all the I/O stuff, system registers and so-forth that is really 11/780-specific. 'course in '77 there was only the 11/780; the Comet (11/750) was very much in the future, but the '79 rearrangement suggests that the VAX family idea was now a part of DEC policy.) -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 06 Dec 98 11:40:48 GMT Organization: UltraNet Communications, Inc. Lines: 30 Message-ID: <74drqe$rah$1@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36691051.D3041D1C@tnglwood.demon.co.uk> <74balk$anb$3@ligarius.ultra.net> <36694157.59552017@tnglwood.demon.co.uk> NNTP-Posting-Host: d7.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 6 Dec 1998 12:03:26 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d7 In article <36694157.59552017@tnglwood.demon.co.uk>, Robert Billing wrote: >jmfbahciv@aol.com wrote: > >> I sure am one of them. I encountered DOS in the late 70s but used >> RSXes (many flavors of it) in the early 70s. When was DOS developed? > > Well, it was relatively new and at V8.2 when I used it in my first paid >programming job on a PDP/11-40 in 1973. I would suspect that the >original release was around 1971. Major releases after this were 8.8, >9.0 and 10.x which was then replaced by 10. Those version numbers are striking a bell but that bell is way, way far away :-). > > It supported asynchronous I/O, and various multi-tasking kernels were >built on top of it for ad-hoc systems, until it was finally replaced by >RSX. > I still don't remember. What was its monitor mode prompt? I did all of the on-line documentation in 1971 and 1972 in addition to typing in the programmers' code (those -11 and -8 guys didn't type then, especially the diagnotic types). Are you sure it was a PDP-11 OS? /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 06 Dec 98 11:43:19 GMT Organization: UltraNet Communications, Inc. Lines: 18 Message-ID: <74drv4$rah$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <74bahs$anb$2@ligarius.ultra.net> <3669420B.F36042F4@tnglwood.demon.co.uk> NNTP-Posting-Host: d7.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 6 Dec 1998 12:05:56 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news.ultranet.com!d7 In article <3669420B.F36042F4@tnglwood.demon.co.uk>, Robert Billing wrote: >jmfbahciv@aol.com wrote: > >> Or, if the program requested it, returns control to the program to >> do other stuff; we called it non-blocking I/O (which was real >> dangerous to do), buffered mode I/O which allowed the program > > Whaddyamean dangerous? Real Programmers always do non-blocking I/O, it >lets you get more done that way. > Shhhh!!!!! Don't tell the children. :-) /BAH Subtract a hundred and four for e-mail. ###### Newsgroups: alt.folklore.computers From: kaplow_r@eisner.decus.org (Bob Kaplow) Subject: Re: Is VMS taboo, or what? Lines: 36 Organization: Northern Illinois Rocketry Assn Message-ID: <1998Dec6.080834.1@eisner> References: <73ela0$51h@hermes.acs.unt.edu> <73eokf$16lk$1@nntp1.u.washington.edu> <73rfg5$dku$4@strato.ultra.net> X-Trace: news.decus.org 912949717 3305 KAPLOW_R [192.67.173.2] X-Nntp-Posting-Host: eisner.decus.org Date: Sun, 6 Dec 1998 13:08:34 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!netnews.com!newspeer1.nac.net!news.new-york.net!news.decus.org!eisner!kaplow_r In article <73rfg5$dku$4@strato.ultra.net>, jmfbahciv@aol.com writes: > In article , > alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >>In article <73g0uo$o0k$1@supernews.com> "Jack Peacock" > >>writes: >> >>>First, let's not forget DECnet goes back quite a way, mid 70's, before >>>Ethernet even got started. >> >>Ethernet dates from around 1971 or 1972, as does the first >>version of DECNET. > > I don't remember it being that early. What I do remember is > that those DECNET folks kept dithering on about the specs for > so long that the TOPS10 folks got impatient and wrote their > own--it's called ANF-10. DECNET didn't implement all the > stuff that was in ANF-10 until (I think) Phase IV. The ANF-10 > developers would go to 198x DECNET review meetings, mention in an > offhand way, "Oh, yea, we did that in 1975", to the great disbelief > of the VMSers. > > There...got that off my chest :-). DECnet predates VMS by many years, to the early 70s IIRC. DECnet was available under RSX, RSTS, RT-11, and I even recall a version for the PDP-8, probably under RTS-8 although I don't recall ever actually using it. Ethernet grew out of an RF implementation to tie universities together in Hawaii, hence the name. Xerox, later joined by DEC and Intel made it into the LAN it became in the early 80s. DEC and DECnet were among the earliest adopters of ethernet. Bob Kaplow ex-dec 1980-1991 ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 06 Dec 1998 13:29:08 GMT Organization: . Lines: 30 Message-ID: <366f8649.5929305@news.uunet.be> References: <366a6629.4860679@news.uunet.be> <74buh4$sd@axalotl.demon.co.uk> NNTP-Posting-Host: pool02b-194-7-150-240.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach huge@nospam.demon.co.uk (Hugh Davies) on 5 Dec 1998 18:37:24 GMT to alt.folklore.computers: > In article <366a6629.4860679@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) writes: > > >> Which RSX did you use? > > > >Sorry, but here comes another stupid question: > >Was there more than one? (slightly embarrassed emoticon here) > > He-he-he. > > -S, -M, -M+ & -D, that I know of. Ahhh - that rings a bell (please excuse me for having forgotten about those suffixes somewhere between 1980-81 and now). It was -M. BTW, this thread made me decide to go searching for some more info on the web - if you're interested in the PDP-11 family but don't have the hardware, take a peep at http://www.dbit.com/ ("unlimited personal/hobby use" on that page sounds interesting) It didn't look as if it contains an OS too, but I'm downloading the docs while writing this, so I guess I'll soon find out. -- "Managing senior programmers is like herding cats." --Dave Platt ###### Message-ID: <36690D71.4F611118@trailing-edge.com> From: Tim Shoppa Organization: Trailing Edge Technology X-Mailer: Mozilla 3.03Gold (X11; I; OpenVMS V7.0 DEC 3000 Model 300L) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 129 Date: Sat, 05 Dec 1998 10:39:45 -0400 NNTP-Posting-Host: 198.232.144.27 X-Trace: audrey2.cais.com 912872935 198.232.144.27 (Sat, 05 Dec 1998 10:48:55 EDT) NNTP-Posting-Date: Sat, 05 Dec 1998 10:48:55 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!in1.nntp.cais.net!199.0.216.204.MISMATCH!audrey2.cais.com!not-for-mail Sergej Roytman wrote: > > Of course, VMS beginners from other backgrounds may find > >commands like "dir" (or "directory") to list a directory or "copy" to > >copy files not *that* strange, perhaps not even "run" to run executable > >programs (necessary if they are not installed as commands). > > My response was: what's the connection with MessyDOS? Hm, since this > is the folklore newsgroup, what _is_ the connection? At a guess, > > CP/M->QDOS->The Program-Loader of the Beast? CP/M, MS-DOS, VMS, RSX, RT-11, and many other OS's are all directly traceable, in their command syntax, back to OS/8 (a PDP-8 operating system) which itself has parts of its command syntax traceable back to the PDP-6 monitor and, before that, the PDP-1 monitor. With many holes - especially in the Pre-OS/8 lineage - I'd draw this sort of diagram (proportional fonts off, everybody!!: /->RSTS->RSTS/E-------\ | | /->DOS-11-|->RT-11--------------| | | | | \->RSX-11D-->RSX-11M+-| | | | \ | | \->IAS-------->VMS PDP-1-->PDP-6--->OS/8-| | monitor monitor | \ \---------->CP/M-->MS-DOS Note that, unlike many others, I tend to think that VMS is most closely related to IAS in "core structure", though clearly it had many RSTS/E, RSX-11M+, and RT-11-type features layered over it. In the late 70's/early 80's, there was an effort inside DEC to unify the RSTS/E, RT-11, and RSX-11 command lines through the introduction of CCL (Concise Command Language)/DCL (Digital Command Language) syntaxes. The fact that this was done from "top-down" is evident in the extreme consistency between qualifiers - for example, wherever "/BEFORE" makes sense to select a subset of files, it's done consistently. Contrast this with Unix, where the command syntax evolved from the "bottom", and didn't make its way up at all. For instance, look at the way the "-r" qualifier is treated by different Unix commands: rm: "-r" means to delete files recursively on down subdirectories. ls: "-r" reverses the order of an alphabetic or date sort mail: "-r" prints mail messages in first-in, first-out order man: "-r" forces a reformat of the manual page, but do not display it who: "-r" indicates the current run-level of the init process Many CP/M "file-system" ideas are obviously straight descendants from the DOS-11 file system. (Of course, the UIC structure was greatly abbreviated into "user numbers".) The CP/M command syntax is closer to OS/8, IMHO. > >In lots of ways VMS is very consistent, and in lots of ways it is very > >flexible. In the late 70's/early 80's, there was an effort inside DEC to unify the RSTS/E, RT-11, and RSX-11 command lines through the introduction of CCL (Concise Command Language)/DCL (Digital Command Language) syntaxes. The fact that this was done from "top-down" is evident in the extreme consistency between qualifiers - for example, wherever "/BEFORE" makes sense to select a subset of files, it's done consistently. > >These are, though, very different from the consistent (where > >existent) and flexible ways of Unix. Where is the command syntax of Unix *ever* consistent, except by random chance? For example, contrast the consistency of DCL options options with the command-line options for some very common Unix /bin commands (here I chose SunOS - a different manufacturer's Unix is very likely to use an entirely different set of options): rm: "-r" means to delete files recursively on down subdirectories. ls: "-r" reverses the order of an alphabetic or date sort mail: "-r" prints mail messages in first-in, first-out order man: "-r" forces a reformat of the manual page, but do not display it who: "-r" indicates the current run-level of the init process As you can see, usually "-r" means something completely different between different commands. In the case of "ls" and "mail", where it has something to do with sorting items in a list, it means a different order for "ls" than for "mail"!!! > >This is the reason that Unix people > >very often find VMS strange and develop an intense dislike towards VMS. On the other hand, VMS people develop an intense dislike for Unix, finding fatal design flaws in the command interface almost immediately. For example, my first experience with Unix was to compile and run a simple C program. I called my first attempt at a C program "test.c". After a couple of iterations with "cc" and putting semicolons in the right places, I had an executable called "test" in my current directory that was supposed to do a couple simple calculations and display the result. So I type "test" to run my new program - and nothing comes out. I look around to see if maybe it wrote its output to some file in my directory, but there's nothing there. So we start calling in other physicists who've used Unix in the past to look at the code and figure out what I'm doing wrong. Strange thing is, some of them can put "test.c" in their home directory and produce an identical "test" executable which *does* do the calculations and produce the results when they type "test". Of course, eventually I learned my first "school of hard knocks" lesson from Unix, and the necessity of the "which" command, and how it relates to the PATH environment variable. But it took us several hours to figure this out. I can't think of a single time under VMS where it took me, and a bunch of Ph.D. nuclear phycisists, more than a couple minutes to type in a ten-line program and get the results. But Unix brought an obvious decline in productivity beginning from day one! Tim. (shoppa@trailing-edge.com) ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 06 Dec 1998 20:09:14 -0600 Organization: MindSpring Enterprises Lines: 10 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> NNTP-Posting-Host: d1.56.d2.2c X-Server-Date: 7 Dec 1998 02:05:57 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk>, dg@ (David Given) wrote: >Hey! How about an open source VMS clone based around the Linux kernel? Fer God's sake, why? The VMS kernel is the -good- part. Now, a (real) OpenVMS... w/ source code and no licenses.... You're just griping about the command set, not the OS. -- Howard S Shubs The Denim Adept ###### From: dg@ (David Given) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 06 Dec 1998 23:26:27 GMT Organization: I'm organised? Wow! Message-ID: <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: taos.demon.co.uk X-NNTP-Posting-Host: taos.demon.co.uk:158.152.120.224 X-Trace: news.demon.co.uk 912986787 nnrp-01:12615 NO-IDENT taos.demon.co.uk:158.152.120.224 X-Complaints-To: abuse@demon.net Lines: 51 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!diablo.theplanet.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!taos.demon.co.uk!!dg In article <749lip$gjf$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: >In article <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk>, >Edward John M. Brocklesby wrote: >>On about Thu, 3 Dec 1998 00:02:00 +1030, The XO might have written: >>[snip story about HD] >> >>I have a Linux box here that has a single 8.5Gb EIDE HD running the whole >>thing. Occasionally, I do something stu^H^H^Hinteresting, like drop it, >>or turn it off accidently (it's on a seperate PSU to the computer), and >>whatever I do, Linux manages to recover from it fine. (Unlike NetBSD, >>which just hangs when I turn the HD off). >> >>That's on Amiga hardware, don't know if it would work on PC's. > >I tried doing this to my Linux machine here (PC), I pulled out the hard >drive cable, did some file I/O, replugged it in, and I saw one of the >most spectacular Linux crashes ever. All of the windows (displaying >with X remotely to my Windows workstation) starting closing one by one, >my telnet sessions to the machine started closing, and the console >completely locked up. Part of the reason is the particularly lousy PC I/O protocols (I assume you're using IDE here, SCSI's rather better). What *should* happen is that when you take the hard-drive off-line, any processes with I/O transactions pending (including swap) should block and everything else should keep running. When you plug it back in the transactions should complete and the processes unblock. This requires a reasonably secure I/O protocol between the computer and the drive. I believe Amiga hardware requires smart drives, which would explain why that still ran. Of course, another part of the reason is that its not been an issue people have been particularly bothered about fixing, seeing as most developers use PC hardware. >Of course, nobody expected Linux to be as robust as VMS, did they? ;-) No reason why it shouldn't be. VMS has a lot to teach the Unix community (some of it good examples of things not to do, like the way it implemenets have the OS functionality in the SET command and the other half in the SHOW command). Hey! How about an open source VMS clone based around the Linux kernel? -- +- David Given ------------McQ-+ | Work: dg@tao.co.uk | "Evil will triumph because good is | Play: dgiven@iname.com | DUMB!" --- Dark Helmet +- http://wiredsoc.ml.org/~dg -+ ###### From: tangentSPAMCATCHER@cyberport.com (Warren Young) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 02:24:32 GMT Organization: none Lines: 39 Message-ID: <366b39eb.3006571501@news.cyberport.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> NNTP-Posting-Host: 56k12-118.cyberport.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: macaw.cyberport.com 912997478 18136 204.134.118.12 (7 Dec 1998 02:24:38 GMT) X-Complaints-To: abuse@cyberport.com NNTP-Posting-Date: 7 Dec 1998 02:24:38 GMT X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!news-xfer.newsread.com!netaxs.com!newsread.com!news.winstar.net!winstar!news.good.net!set!news.cyberport.com!not-for-mail ftit@engin.umich.edu (Sergej Roytman) wrote: >>>> >The post before said that VMS managed even with the VM/paging disk out >>>> >of action. How? >>>> Sigh! It's called timesharing. That's an aspect of implementation >>>Sorry if this is a stupid remark, but I don't see in what way the >>>fact that it's a timesharing system can have anything to do with >>>being able to continue whatever it's doing when the system disk >>>goes off-line (or after it comes back on)? >>_Proper implementation_ of timesharing > >For those yunguns among us who think Unix is pretty cool (well, it >is!), please clarify. I mean, I can see things working OK if >everything that's needed from the disk is in cache, but what happens >the first time the OS decides it's time to page something out? I've only been working with computers seriously since about 1990 (and a nerdy little whiz kid since 1985), but it seems to me that you could do this with a preemptible, multitasking kernel. No, not a kernel that provides multitasking to the user programs, but that can be preempted itself. This is probably easiest to do with a microkernel-style implementation. I know NT has a semi-micro-kernel setup, and I suspect that VMS does, too. So, when the swapper or pager process decides it needs to move a block of memory to the swap file to free up some memory, it simply calls the disk I/O routines to do this. They then realize that the volume in question is apparently unavailable. Now, since it hasn't been formally dismounted, it's probably a temporary error, so the call blocks. The pager then sits there, blocking on a disk I/O call until it completes. Alternately, that disk I/O call could complete after passing the data off to the cache subsystem. The cache subsystem would then wait patiently for the disk to become available again. A kernel like Linux's (which I've been studying recently) _might_ be able to do this, but probably can't at the moment. It's just not designed for it. ###### Newsgroups: alt.folklore.computers From: bdb@GTS.Net (Hello Kittyhawk) Subject: Re: Is VMS taboo, or what? Message-ID: Organization: G.T.S., Toronto, Ontario X-Newsreader: trn 4.0-test66 (4 June 1998) References: <73ela0$51h@hermes.acs.unt.edu> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> Date: Mon, 7 Dec 1998 04:52:55 GMT Lines: 14 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.nyu.edu!cyclone.news.idirect.com!island.idirect.com!tor-nx1.netcom.ca!gts!bdb In article <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk>, David Given wrote: | |Hey! How about an open source VMS clone based around the Linux kernel? How about an open source Linux clone based around the VMS kernel? -- ,u, Bruce Becker Toronto, Ontario 1 416 699 1868 a \i\ Internet: bdb@gts.org Uucp: ...!gts!bdb `/o/-e "Police in helicopter, you keep burnin' down the herb, _\ >_ we going to burn down the cane crop" - for Jeff ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 06:55:18 +0000 Organization: Tanglewood Message-ID: <366B7BD6.F3D24F1C@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36691051.D3041D1C@tnglwood.demon.co.uk> <74balk$anb$3@ligarius.ultra.net> <36694157.59552017@tnglwood.demon.co.uk> <74drqe$rah$1@ligarius.ultra.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913016273 nnrp-01:20570 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 21 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail jmfbahciv@aol.com wrote: > I still don't remember. What was its monitor mode prompt? I did The prompts were $ Normal monitor . Monitor after a ctrl/c (program suspended) # Application * Data entry or editor commands > type then, especially the diagnotic types). Are you sure it > was a PDP-11 OS? Sure? I a) used it from 1973 to 1978 and b) still have the manuals. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 07:00:10 +0000 Organization: Tanglewood Message-ID: <366B7CFA.2A976248@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913016274 nnrp-01:20570 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail David Given wrote: > Hey! How about an open source VMS clone based around the Linux kernel? What a good idea. I have been considering writing a DCL-like shell for Linux. Seriously though, the one point on which I think VMS scores over Linux is that globbing is done in the application, not the shell, so that you can do "related filename" tricks, such as copying a directory tree with the copy command. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 07:12:11 +0000 Organization: Tanglewood Message-ID: <366B7FCB.4EF50E71@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <74b9ad$anb$1@ligarius.ultra.net> <366a6629.4860679@news.uunet.be> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913016275 nnrp-01:20570 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 31 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Luc Van der Veken wrote: > Fortran compiler and run something called TKB - the 'task > builder', which I suspected to be something like a linker. And > run the compiled program. Linker and loader. The normal output from TKB is a file consisting of label blocks, followed by an assortment of binary images of the various overlays of the program. The joy of this is that the .tsk file is a bit-for-bit image of what goes in core, so tasks can be loaded with a single DMA operation, and the pure areas don't have to be swapped as the task file contains what would have been in the swap file. The side benefit of this is that a task could be "installed" with the ins command, which created a tcb, which contained a pointer to the binary on disk. This task would then sit dormant until required, then fire up without any directory searching when required as all RSX had to do was DMA in te code and start it. If you installed a task with a name like ...xyz, then typing xyz would cerate a clone tcb called xyztnn where tnn is the name of your terminal. This tcb would share pure areas with other instances of the task, so that if several people, for example, typed mac to run the macro assembler, only one copy of the code, but several of the variables, would sit in core. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 10:07:15 -0600 Organization: MindSpring Enterprises Lines: 70 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: d1.56.d0.03 X-Server-Date: 7 Dec 1998 16:07:19 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <36690D71.4F611118@trailing-edge.com>, Tim Shoppa wrote: >On the other hand, VMS people develop an intense dislike for Unix, >finding fatal design flaws in the command interface almost immediately. Yes, exactly. There's no -sense- to UNIX; no unity at all. There may be something nice in the kernel design, but since I never see that, it's academic (literally). What -I- see in UNIX: 1) poor command-line design 2) lack of developer support other than the "make" command 3) lack of portability (1) was explained by the message I'm replying to, but I've got an additional issue: documentation. When you have a problem, what do you ask for? "MAN! MAN! I need MAN!" Well, maybe you do if you're female or gay, but the rest of us ask for HELP! Maybe there's some other language where "man" == "help", but it's not one a normal user will think of. The -first- thing a new user needs, other than to know how to log in, is how to get to the documentation. How is that user to find out how to do that? They won't know to use "man" off-hand. It's about as unintuitive a command as I can think of, not that I can't think of others: grep (mister, when I want to search something, I don't think "GREP! Yeah, GREP that puppy" unless I've got experience with UNIX. Specifically.), ls (list! Yeah! Now -of course- you take the "i" and the "t" out. That's obvious, right?), dbx (okay, you can argue that a programmer has cause to get more information before they first log in, but still.), and I could go on for almost every command I've ever seen on UNIX. (2) Let me ask you about the UNIX source-level debuggers. So, if you get a crash, you can use the "core" file. Great. What if you want to find something being changed inapproprately? A watchpoint takes FOREVER, and in my experience, usually CRASHES THE DEBUGGER after taking ALL NIGHT to find what I've been looking for. The debuggers under UNIX make me go back to pre-debugger technologies of "write value of variable at various points in the program". Yes, that's advanced. How much cooperation do you get on UNIX between languages? Okay, I can do sorta kinda okay (Contortions 'R' Us) using f77 and cc, but if I wanted to use other language, what new obscurities would I need to discover? UNIX has no calling standard. It has none of the benefits such a standard would bring, such as ease of development due to better debuggers and ease of using different languages for different parts of a program when appropriate. UNIX has no consistancy of library routines between languages. If I want to do everything in C, I'm golden, but not everyone enjoys programming in assembly language all the time. If I try to call system routines or RTL routines from f77, I've got to pray there's some kind of wrapper available, or WRITE IT MYSELF in C, and do my contortions to make it all link properly. (3) So, if I take a binary from UNIX box X to UNIX box Y, it'll just -work-, right? Y'know, like I can take a VMS binary from a 20-year-old VAX and run it on the newest VAX. Ditto with VMS on an Alpha. Okay, if you disallow that because they both come from the same vendor, how about language compilers at least? If I build arbitrary program A on an HP/UX box, will I be able to build it WITHOUT ANY CHANGES on AIX or AUX or Digital UNIX? UNIX... no wonder it's being supplanted by Windows NT. And the fact that VMS appears to be ported -from- more often than it's ported -to-, is nothing less than a major tragedy. -- Howard S Shubs The Denim Adept ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 98 12:13:25 GMT Organization: UltraNet Communications, Inc. Lines: 26 Message-ID: <74gi3p$k6s$1@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36691051.D3041D1C@tnglwood.demon.co.uk> <74balk$anb$3@ligarius.ultra.net> <36694157.59552017@tnglwood.demon.co.uk> <74drqe$rah$1@ligarius.ultra.net> <366B7BD6.F3D24F1C@tnglwood.demon.co.uk> NNTP-Posting-Host: d11.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 7 Dec 1998 12:36:09 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!netnews.com!newsswitch.lcs.mit.edu!news.ultranet.com!d11 In article <366B7BD6.F3D24F1C@tnglwood.demon.co.uk>, Robert Billing wrote: >jmfbahciv@aol.com wrote: > >> I still don't remember. What was its monitor mode prompt? I did > > The prompts were > > $ Normal monitor > . Monitor after a ctrl/c (program suspended) > # Application > * Data entry or editor commands > >> type then, especially the diagnotic types). Are you sure it >> was a PDP-11 OS? > > Sure? I a) used it from 1973 to 1978 and b) still have the manuals. > Chuckle. That's the neat thing about computers...just when you think you've seen it all someone comes around to remind one that s/he hasn't. :-) /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 98 13:11:46 GMT Organization: UltraNet Communications, Inc. Lines: 69 Message-ID: <74glh7$j1n$1@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> NNTP-Posting-Host: d11.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 7 Dec 1998 13:34:31 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!netnews.com!feed1.news.rcn.net!rcn!news.ultranet.com!d11 [somebody's hacked off the atributes and it tweren't me] In article <366b39eb.3006571501@news.cyberport.com>, tangentSPAMCATCHER@cyberport.com (Warren Young) wrote: >ftit@engin.umich.edu (Sergej Roytman) wrote: > >>>>> >The post before said that VMS managed even with the >>>>> >VM/paging disk out of action. How? >>>>> Sigh! It's called timesharing. That's an aspect of implementation >>>>Sorry if this is a stupid remark, but I don't see in what way the >>>>fact that it's a timesharing system can have anything to do with >>>>being able to continue whatever it's doing when the system disk >>>>goes off-line (or after it comes back on)? >>>_Proper implementation_ of timesharing >> >>For those yunguns among us who think Unix is pretty cool (well, it >>is!), please clarify. I mean, I can see things working OK if >>everything that's needed from the disk is in cache, but what happens >>the first time the OS decides it's time to page something out? > >I've only been working with computers seriously since about 1990 (and >a nerdy little whiz kid since 1985), but it seems to me that you could >do this with a preemptible, multitasking kernel. No, not a kernel >that provides multitasking to the user programs, but that can be >preempted itself. This is probably easiest to do with a >microkernel-style implementation. I know NT has a semi-micro-kernel >setup, and I suspect that VMS does, too. You must consider the case where your microkernel preempts itself. > >So, when the swapper or pager process decides it needs to move a block >of memory to the swap file to free up some memory, it simply calls the >disk I/O routines to do this. They then realize that the volume in >question is apparently unavailable. Now, since it hasn't been >formally dismounted, it's probably a temporary error, so the call >blocks. The pager then sits there, blocking on a disk I/O call until >it completes. Assuming it is a temporary error (whatever that means...there are myriads of categories), do you tell someone about the error? And, if you do think someone should be notified, who? > >Alternately, that disk I/O call could complete after passing the data >off to the cache subsystem. The cache subsystem would then wait >patiently for the disk to become available again. Thus you leave the user or program with the understanding that the data has been written in iron. Now the next action by the same user (or, more likely, another user) is to read back the newly written data. How do you ensure data integrity? > >A kernel like Linux's (which I've been studying recently) _might_ be >able to do this, but probably can't at the moment. It's just not >designed for it. Think about the notions I've been mentioning above. Each OS (and I'm talking about decent implmentations) has their own philosophy of the approach taken. If you would like, I'd be willing to discuss overviews. /BAH Subtract a hundred and four for e-mail. ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec7.174534.26644@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <73ela0$51h@hermes.acs.unt.edu> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> Lines: 19 NNTP-Posting-Host: 130.244.238.178 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 913053483 130.244.238.178 (Mon, 07 Dec 1998 18:58:03 MET DST) NNTP-Posting-Date: Mon, 07 Dec 1998 18:58:03 MET DST X-Sender: s-64812@dialup238-12-2.swipnet.se Date: Mon, 7 Dec 1998 17:45:34 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article <366B7CFA.2A976248@tnglwood.demon.co.uk>, Robert Billing wrote: > Linux. Seriously though, the one point on which I think VMS scores over > Linux is that globbing is done in the application, not the shell, so > that you can do "related filename" tricks, such as copying a directory > tree with the copy command. If the globbing is done in the application then implementing file name completion will be difficult. What kind of difficulties related to globbing do you have with Copying a directory tree? If the application needs to see a filename before globbing all you have to do is put single quotes around the glob pattern. The reverse for DCL is not that easy. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 18:44:29 +0000 Organization: Tanglewood Message-ID: <366C220D.2710F87B@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913057250 nnrp-10:8686 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 29 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Goran Larsson wrote: > related to globbing do you have with Copying a directory tree? Try to copy all the *.new files in a tree to *.old in another tree with a different root but the same topology. There may be an easy way to do this but I'm still in the process of converting from VMS to Linux mentally. Personally I prefer VMS to unix, and there are a lot of things wrong with unix, but none of them are fundamental. Replacing the *entire* shell and command set with something DCL-like would not break unix, and done carefully it could co-exist with the existing system. I strongly suspect that Linux will evolve into something vastly more user-friendly over the next 10 years. In fact I was scribbling some notes on how I might contribute to this, only this morning. OTOH there is so much fundamentally wrong with M$win, that there is no hope it its ever being right without starting again. Unix has its problems, but the difference is between a rather shabby, but structurally sound, building, and palatial decor sliding into a swamp. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: dcurry@silo.csci.unt.edu (David Mason Curry) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 18:50:55 GMT Organization: University of North Texas Lines: 33 Message-ID: <74h82f$4mt@hermes.acs.unt.edu> References: <73ela0$51h@hermes.acs.unt.edu> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> NNTP-Posting-Host: silo.csci.unt.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!cs.utexas.edu!news.unt.edu!silo.csci.unt.edu!dcurry Goran Larsson (hoh@approve.se.NO_JUNK_EMAIL) wrote: : In article <366B7CFA.2A976248@tnglwood.demon.co.uk>, : Robert Billing wrote: : > Linux. Seriously though, the one point on which I think VMS scores over : > Linux is that globbing is done in the application, not the shell, so : > that you can do "related filename" tricks, such as copying a directory : > tree with the copy command. : If the globbing is done in the application then implementing : file name completion will be difficult. What kind of difficulties : related to globbing do you have with Copying a directory tree? : If the application needs to see a filename before globbing all you : have to do is put single quotes around the glob pattern. The : reverse for DCL is not that easy. Actually, globbing is better to do in the application, because the application can know what was entered on the command line. Hence, commands like "rename *.foo *.bar" could conceivably work like one would probably like them to. However, "mv *.foo *.bar" is only useful once in a blue moon under the UN*X paradigm and is in any case quite scary because it might surprise you. And without version control, that is often bad. And why would adding globbing to the shell be so much easier than having a library call perform globbing from the application? : -- : Goran Larsson hoh AT approve DOT se : I was an atheist, http://home DOT swipnet DOT se/hoh/ : until I found out I was God. Cheers, David Curry ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 07 Dec 1998 22:32:56 +0100 Organization: My own Private Self Lines: 75 Sender: neil@chonsp.franklin.ch Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> X-Newsreader: Gnus v5.3/Emacs 19.34 dcurry@silo.csci.unt.edu (David Mason Curry) writes: > > Goran Larsson (hoh@approve.se.NO_JUNK_EMAIL) wrote: > : Robert Billing wrote: > > : > Linux. Seriously though, the one point on which I think VMS scores over > : > Linux is that globbing is done in the application, not the shell, so Not just Linux. Globbing in the shell hab been there in all Unixes since the very beginning. It was originally an separate program "glob" (that is where the name comes from), called by the shell, which then called the actual program that used the file names. > : If the globbing is done in the application then implementing > : file name completion will be difficult. Now why would completing an filename be difficult? If I have typed "ver" and there is only one file "very_long_named_file" beginning with "ver" the shell will expand. What has globbing (*,?) got to do with this? > : If the application needs to see a filename before globbing all you > : have to do is put single quotes around the glob pattern. Increases typing to put "" aroung every globbed filename. And having all programs expect pre-globbed input and therefore not glob it themselves makes switching it off a big downer. > commands like "rename *.foo *.bar" could conceivably work like one > would probably like them to. However, "mv *.foo *.bar" is only > useful once in a blue moon under the UN*X paradigm and is in any > case quite scary because it might surprise you. And it is usually very suicidal for your poor files. And then there is also: rm * .o can not be caught and ask "do you really want to delete all? rm * with an file -r in directory kills all subdirectories (and -r stays!) rm -this-is-a-file can not be tricked with *-a-file, needs rm - -this* rm "file name with spaces in it *" fails to expand rm 'file name with spaces in it *' expands and fails, no "file" "name" ... rsh cat file* needs "" to make it glob on remote system rcp "remote:files*" . also needs "" to stop local globbing grep files_* /dev/null to prevent 1-file behaviour (no file_name: before line) no standard to have an -? option for help (? is an 1 character glob) Shell globbing is IMHO the most losing feature in Unix. > And without version control, that is often bad. That is something I have a few times thought of adding to Linux. > And why would adding globbing to the > shell be so much easier than having a library call perform globbing > from the application? It isn't easier. It is just simply historically compatible. It was originally done so because Unix had on the pdp-11 no shared libraries. And linking in glob to every program was too disk filling (and it also increased swapping (full processes!) time (originally only one process in the machine)). For an 3 person text processing system it was an acceptable case of scrounching. Today it is an massive annoyance. -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### Message-ID: <366C2A48.353@gazonk.del> Date: Mon, 07 Dec 1998 14:19:36 -0500 From: "Foobar T. Clown" Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 7 Dec 1998 14:20:57 -0500, 198.133.110.49 Lines: 21 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!ix.netcom.com!zeus.nomos.com!198.133.110.49 Howard S Shubs wrote: > > UNIX... no wonder it's being supplanted by Windows NT. At some point, you have to ask yourself, what does "UNIX" mean anyway? Here's what the current owners of the trademark have to say about UNIX and WinDoze NT: Microsoft® Windows NT [...] has no connection with the UNIX system source code. However, market demand for POSIX.1, POSIX.2 has led to developments by several companies of add-ons that provide partial functionality. Should the functionality meet the requirements of the UNIX brand then indeed it could become a UNIX branded system. You can experience the full horror by following the various links from this URL: http://www.UNIX-systems.org/what_is_unix.html Click on the link labelled "What About All Those 'Flavors'" to find the text I quoted above. ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> From: Marco S Hyman Date: 07 Dec 1998 15:20:54 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 Lines: 58 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 913072855 220 marc@204.94.187.130 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news1.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail hshubs@mindspring.com (Howard S Shubs) writes: > (1) was explained by the message I'm replying to, but I've got an > additional issue: documentation. When you have a problem, what do you ask > for? "MAN! MAN! I need MAN!" Well, maybe you do if you're female or > gay, but the rest of us ask for HELP! Maybe there's some other language Hmmm, on my unix-like system when I type in "help" I get... MAN(1) OpenBSD Reference Manual MAN(1) NAME man - display the on-line manual pages SYNOPSIS man [-achw] [-C file] [-M path] [-m path] [-s section] [-S subsection] [section] name [...] man -k keyword man -f filename ... > (3) So, if I take a binary from UNIX box X to UNIX box Y, it'll just > -work-, right? Y'know, like I can take a VMS binary from a 20-year-old > VAX and run it on the newest VAX. Ditto with VMS on an Alpha. Sometimes... example: [My OS] $ uname -ms OpenBSD i386 [The OS netscape is compiled for] $ file netscape netscape: BSD/OS i386 compact demand paged executable It runs fins. Agreed, both systems are BSD based and they are both for the same architecture. No, it wont run on my Sun. Curious, will your VAX binary run on the alpha (I'd guess yes)? How about the other direction? > Okay, if you disallow that because they both come from the same vendor, > how about language compilers at least? If I build arbitrary program A on > an HP/UX box, will I be able to build it WITHOUT ANY CHANGES on AIX or AUX > or Digital UNIX? Probably not (without some extra effort on the part of the original developers). But I'd guess the effort is not beyond reason as there are many programs out there that can be compiled and installed on different systems (sometimes including VMS and WinNT) using steps as simple as $ configure $ make $ make install // marc ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 20:22:22 GMT Lines: 23 Message-ID: <74hddu$ukl$3@irk.zetnet.co.uk> References: <74drv4$rah$2@ligarius.ultra.net> NNTP-Posting-Host: man-043.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 913062142 31381 194.247.41.53 (7 Dec 1998 20:22:22 GMT) NNTP-Posting-Date: 7 Dec 1998 20:22:22 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-06 jmfbahciv@aol.com said: :Robert Billing wrote: :> Whaddyamean dangerous? Real Programmers always do non-blocking :>I/O, it lets you get more done that way. :Shhhh!!!!! Don't tell the children. :-) The children already know, so there. It's just that the commonest hardware is the IBM PC, and I believe that IDE drives are handled in such a way as to make non-blocking disk reads virtually impossible to implement. The best you can do is queue reads and then execute them in order when you've got a reasonable sweep. Someone please tell me if I'm wrong (and a quick description of how would also be nice ;> ) but I've been brought up to believe that the AT and above use the INSB instruction to do 512 reads from a port per sector, and getting them to do it with DMA is a bit of a nightmare. And certainly the PC BIOS blocks on read. Woof. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: hshubs@mindspring.com's message of Mon, 07 Dec 1998 10:07:15 -0600 Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> Date: Mon, 7 Dec 1998 20:33:47 GMT Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!netnews.com!newsfeed.enteract.com!ix.netcom.com!netcom!alderson In article hshubs@mindspring.com (Howard S Shubs) writes: >It's about as unintuitive a command as I can think of, not that I can't think >of others: [snip] >ls (list! Yeah! Now -of course- you take the "i" and the "t" out. That's >obvious, right?) Well, "ls" is short for "list segments" and derives from the Multics command of the same name... -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: Robert Billing's message of Mon, 07 Dec 1998 07:00:10 +0000 Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> Date: Mon, 7 Dec 1998 20:42:37 GMT Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!nntp.abs.net!newsfeed.enteract.com!ix.netcom.com!netcom!alderson In article <366B7CFA.2A976248@tnglwood.demon.co.uk> Robert Billing writes: >Seriously though, the one point on which I think VMS scores over Linux is >that globbing is done in the application, not the shell, so that you can >do "related filename" tricks, such as copying a directory tree with the >copy command. Is globbing done *in the application*, or is it done by the operating system via a (set of) system call(s) as in Tops-20? I realize the 32-bit weenies had a serious case of NIH, but did they really get that wrong, too? -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: mww@microfocus.com's message of 5 Dec 1998 05:11:41 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <749l84$7vd$1@agate.berkeley.edu> <74afad$nk8@news1.newsguy.com> Date: Mon, 7 Dec 1998 21:00:40 GMT Lines: 27 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!ix.netcom.com!netcom!alderson In article <74afad$nk8@news1.newsguy.com> mww@microfocus.com (Michael Wojcik) writes: >In article <749l84$7vd$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) writes: >>I can't think of any reason for a real OS to crash just because it can't >>reach the page disk. >Unless it has a pageable kernel, like AIX versions 3 and up, and the disk >fails when some crucial portion of the kernel is paged out. (Though of course >that's a fatal page in failure, not a problem with paging out.) Tops-20, and its TENEX predecessor, had swappable and resident portions of the monitor. One of the folkloric aspects of the first version of TENEX is the story that the pager code was not originally declared to be resident, leading to the console error message ?PAGE FAULT IN PAGE FAULT HANDLER (or similar wording--I was *not* there, ok?) I think the Tops-10 monitor was all resident code--but there are those who will correct me if I'm wrong, you know. -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 21:21:52 -0600 Organization: MindSpring Enterprises Lines: 44 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: d1.56.d2.09 X-Server-Date: 8 Dec 1998 03:21:54 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!192.26.210.166!sunqbc.risq.qc.ca!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article , Marco S Hyman wrote: >hshubs@mindspring.com (Howard S Shubs) writes: > >> (1) was explained by the message I'm replying to, but I've got an >> additional issue: documentation. When you have a problem, what do you ask >> for? "MAN! MAN! I need MAN!" Well, maybe you do if you're female or >> gay, but the rest of us ask for HELP! Maybe there's some other language > >Hmmm, on my unix-like system when I type in "help" I get... Does straight UNIX do that? And, does HELP work on its own, w/o referring to MAN? The UNIX command set is waaaaay arbitrary and counter-intuitive. At least DCL started-as/is a command set with some coordination. No one even tries to do that with UNIX. It's an OS built by hackers, and it's mainly useful to hackers. I don't mind that, but when it's compared to an -organized- operating system like VMS and the words describing the shortcomings of UNIX are redefined to make them "good", I have trouble with that. Case-sensitivity as a GOOD thing? Short, arbitrary combinations of letters for normal commands?? Horseshit. All that should be redone to make it easier to use. Keep the older commands for compatibility's sake, sure, but start working on fixing the problems. >It runs fins. Agreed, both systems are BSD based and they are both >for the same architecture. No, it wont run on my Sun. Curious, will >your VAX binary run on the alpha (I'd guess yes)? How about the other >direction? Neither, but a recompile/relink sequence will almost always succeed. >Probably not (without some extra effort on the part of the original >developers). But I'd guess the effort is not beyond reason as there >are many programs out there that can be compiled and installed >on different systems (sometimes including VMS and WinNT) using >steps as simple as > > $ configure An awful lot of effort goes into that one step, for just the reason I mention. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 21:22:49 -0600 Organization: MindSpring Enterprises Lines: 10 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: d1.56.d2.09 X-Server-Date: 8 Dec 1998 03:22:50 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article , alderson@netcom.com wrote: >Well, "ls" is short for "list segments" and derives from the Multics command of >the same name... Interesting. Not useful to an Accounting user, or anyone other than a computer scientist, but interesting. So a "segment" was what we call a "file"? -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 21:25:17 -0600 Organization: MindSpring Enterprises Lines: 19 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <366C2A48.353@gazonk.del> NNTP-Posting-Host: d1.56.d2.09 X-Server-Date: 8 Dec 1998 03:25:18 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!feeder.qis.net!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <366C2A48.353@gazonk.del>, "Foobar T. Clown" wrote: >You can experience the full horror by following the various links from >this URL: > http://www.UNIX-systems.org/what_is_unix.html > >Click on the link labelled "What About All Those 'Flavors'" to find the >text I quoted above. Yeah, they tried to do the same with VMS. It sank like a lead brick. I'll admit that pipes are nice. Last I looked, I think DEC (was DEC at the time) was adding them to VMS, but I'm not certain of that. Been doing UNIX for the last year+. Most of what I like about UNIX is in the shell, though. Perhaps porting the Korn shell to VMS would be a useful project. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 21:28:34 -0600 Organization: MindSpring Enterprises Lines: 17 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> NNTP-Posting-Host: d1.56.d2.09 X-Server-Date: 8 Dec 1998 03:28:34 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <74hp72$lk0$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >In the VMS version you haven't told it which file versions to execute the >command on. I can't remember whether it would parse "[...]*.c" as >"[...]*.c;*" or not. It's been too many years since I used VMS. That'll automatically get it to scan the most-recent (highest) version of each file. >Can anyone really say they like VMS's "not really a hierarchical file system" >structure? How is it not hierarchical? -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 07 Dec 1998 21:35:38 -0600 Organization: MindSpring Enterprises Lines: 10 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> NNTP-Posting-Host: d1.56.d2.09 X-Server-Date: 8 Dec 1998 03:35:39 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!europa.clark.net!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article , alderson@netcom.com wrote: >Is globbing done *in the application*, or is it done by the operating system >via a (set of) system call(s) as in Tops-20? I realize the 32-bit weenies had >a serious case of NIH, but did they really get that wrong, too? Okay, rephrase: "by" the application calling the OS. So it's really happening in the RTL. -- Howard S Shubs The Denim Adept ###### From: postmaster@localhost.edu (Mr. Damon Kelly) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 22:03:26 -0500 Organization: University of Maryland, Baltimore County Lines: 24 Message-ID: <74i4tu$bv@umbc7.umbc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> Reply-To: postmaster@localhost.edu NNTP-Posting-Host: umbc7.umbc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!cpk-news-feed4.bbnplanet.com!cpk-news-feed1.bbnplanet.com!news.gtei.net!news.ums.edu!news.umbc.edu!not-for-mail In article <74hp72$lk0$1@agate.berkeley.edu>, Eric J. Korpela wrote: > >In the VMS version you haven't told it which file versions to execute the >command on. I can't remember whether it would parse "[...]*.c" as >"[...]*.c;*" or not. It's been too many years since I used VMS. VMS, by default, will use only the file with the highest version number. >Can anyone really say they like VMS's "not really a hierarchical file system" >structure? In what ways? If a "fresh" user with no input from others didn't know enough to use SHOW DEVICE/MOUNTED to find other disk devices, life with VMS can be *slightly* more difficult; otherwise, VMS's filesystem is as hierarchial as any other. -d -- Homepage on a machine called, "research," at the University of Maryland, Baltimore County. My username is damon. You know what to do next. ** Note: e-mailbox closed to non-UMBC accounts and most domains until I-net spam is a badly-remembered dream. Replies made on a case-by-case basis. ** ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 22:24:16 GMT Organization: Daedalus Consulting Lines: 49 Message-ID: <74hkig$oaq$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <1998Dec7.174534.26644@lorelei.approve.se>, Goran Larsson wrote: >If the globbing is done in the application then implementing >file name completion will be difficult. What kind of difficulties I don't know what you mean by file name completion being difficult. Having implemented DCL filename completion, believe me, the completion bit was easy. (The bit to get the completion routine actually called by DCL was a *wee* bit trickier... 8-) >related to globbing do you have with Copying a directory tree? >If the application needs to see a filename before globbing all you >have to do is put single quotes around the glob pattern. The >reverse for DCL is not that easy. Shell globbing isn't that great. Shell globbing means all you need to process a globbed list is: for(i = 1; i < argv; i++) process(argv[i]); by contrast, using LIB$FIND_FILE: for(i = 1; i < argv; i++) while(file = find_file(argv[i])) process(file); ("find_file" is a (fictional) wrapper to lib$find_file -- it'd be about half a dozen lines purely to convert VMS style arguments and status returns to something more C-ish. It's hardly VMS's fault that it was designed with other languages in mind, but VMS calls are a little clunky when called from C which has no idea what a string descriptor is.) On the flip side, Unix shell globbing means that the entire file list has to be sucked into memory by the shell, and handed bodily to the subprocess. And the shell doesn't have recursive searches. Thus, with VMS you can mumble: $ command [...]*.c and it'll deal to all .c files in the directory and below, regardless of how many files are involved, while under unix you end up with: % command `find . -name='*.c'` which promptly complains that the argument list is too long and leaves you farting around with the incredible bogosity that is xargs. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 07 Dec 1998 22:32:56 +0100 Organization: My own Private Self Lines: 75 Sender: neil@chonsp.franklin.ch Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> X-Newsreader: Gnus v5.3/Emacs 19.34 dcurry@silo.csci.unt.edu (David Mason Curry) writes: > > Goran Larsson (hoh@approve.se.NO_JUNK_EMAIL) wrote: > : Robert Billing wrote: > > : > Linux. Seriously though, the one point on which I think VMS scores over > : > Linux is that globbing is done in the application, not the shell, so Not just Linux. Globbing in the shell hab been there in all Unixes since the very beginning. It was originally an separate program "glob" (that is where the name comes from), called by the shell, which then called the actual program that used the file names. > : If the globbing is done in the application then implementing > : file name completion will be difficult. Now why would completing an filename be difficult? If I have typed "ver" and there is only one file "very_long_named_file" beginning with "ver" the shell will expand. What has globbing (*,?) got to do with this? > : If the application needs to see a filename before globbing all you > : have to do is put single quotes around the glob pattern. Increases typing to put "" aroung every globbed filename. And having all programs expect pre-globbed input and therefore not glob it themselves makes switching it off a big downer. > commands like "rename *.foo *.bar" could conceivably work like one > would probably like them to. However, "mv *.foo *.bar" is only > useful once in a blue moon under the UN*X paradigm and is in any > case quite scary because it might surprise you. And it is usually very suicidal for your poor files. And then there is also: rm * .o can not be caught and ask "do you really want to delete all? rm * with an file -r in directory kills all subdirectories (and -r stays!) rm -this-is-a-file can not be tricked with *-a-file, needs rm - -this* rm "file name with spaces in it *" fails to expand rm 'file name with spaces in it *' expands and fails, no "file" "name" ... rsh cat file* needs "" to make it glob on remote system rcp "remote:files*" . also needs "" to stop local globbing grep files_* /dev/null to prevent 1-file behaviour (no file_name: before line) no standard to have an -? option for help (? is an 1 character glob) Shell globbing is IMHO the most losing feature in Unix. > And without version control, that is often bad. That is something I have a few times thought of adding to Linux. > And why would adding globbing to the > shell be so much easier than having a library call perform globbing > from the application? It isn't easier. It is just simply historically compatible. It was originally done so because Unix had on the pdp-11 no shared libraries. And linking in glob to every program was too disk filling (and it also increased swapping (full processes!) time (originally only one process in the machine)). For an 3 person text processing system it was an acceptable case of scrounching. Today it is an massive annoyance. -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### From: ftit@engin.umich.edu (Sergej Roytman) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 22:44:14 GMT Organization: University of Michigan, College of Engineering Lines: 22 Message-ID: <74hlnu$1mq@srvr1.engin.umich.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: az.engin.umich.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news-peer.gip.net!news.gsl.net!gip.net!news.maxwell.syr.edu!news.eecs.umich.edu!srvr1.engin.umich.edu!ftit In article <74hkig$oaq$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >In article <1998Dec7.174534.26644@lorelei.approve.se>, >Goran Larsson wrote: > Thus, with >VMS you can mumble: > > $ command [...]*.c > >and it'll deal to all .c files in the directory and below, regardless of >how many files are involved, while under unix you end up with: > > % command `find . -name='*.c'` Huh? % command *.c */*.c works fine for me. -- Sergej Roytman ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: don@news.daedalus.co.nz's message of 7 Dec 1998 22:24:16 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> Date: Mon, 7 Dec 1998 22:48:33 GMT Lines: 31 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!ix.netcom.com!netcom!alderson In article <74hkig$oaq$1@news.wlg.netlink.net.nz> don@news.daedalus.co.nz (Don Stokes) writes: >And the shell doesn't have recursive searches. Thus, with VMS you can mumble: > $ command [...]*.c >and it'll deal to all .c files in the directory and below, regardless of how >many files are involved, while under unix you end up with: > % command `find . -name='*.c'` >which promptly complains that the argument list is too long and leaves >you farting around with the incredible bogosity that is xargs. I've learned to deal with Unix to keep body and soul together when not employed by someone using reasonable operating systems. The right way to do this with the _find_ command is % find . -name '*.c' -exec command {} \; for commands that deal with individual files. For commands that deal with *groups* of files (like _tar_), I'd stick the _find_ output into a temp file and use that in the command. I do wish that the Tops-20 EXEC had a recursive-directory syntax (natively--the GNU utilities do just fine). -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 23:33:27 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 54 Message-ID: <74hok7$ldc$1@agate.berkeley.edu> References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74hddu$ukl$3@irk.zetnet.co.uk>, wrote: >The children already know, so there. It's just that the commonest >hardware is the IBM PC, and I believe that IDE drives are handled in >such a way as to make non-blocking disk reads virtually impossible to >implement. No, they're actually easy to implement on controllers that support them. Because there are controllers that do not support them, Microsoft has chosen to ignore the controllers that do support them. >Someone please tell me if I'm wrong (and a quick description of how >would also be nice ;> ) but I've been brought up to believe that the AT >and above use the INSB instruction to do 512 reads from a port per >sector, and getting them to do it with DMA is a bit of a nightmare. Most modern IDE controllers and drives support two types of I/O, PIO (Polled I/O) and DMA (direct memory access). PIO means the software checks to see if data is ready and then reads it. There are various PIO types PIO-1 through PIO-4 (Maybe higher?) which specify how the computer knows data is ready, how much data will be ready and how wide the transfer is. DMA also comes in several flavors (DMA-1, DMA-2, UltraDMA?). In DMA, the software tells the DMA controller how much data to transfer and where, in memory, to put it. The controller then reports when the transfer is done. Despite what you've heard, DMA is not difficult provided your controller supports it. In practice, non-blocking I/O can be done using both methods. I believe, (but I'm not sure) that some of the PIO methods can be used to generate an interrupt when data is ready. The small size of the PIO transfers means fairly large interrupt overhead, small enough that it may not be worth while to do other things between interrupts. DMA transfers can be larger so if you want to do non-blocking I/O, the DMA method is better. Of course, if you're looking for raw disk I/O performance, busy wait PIO is always fastest, if only for the reason that the RAM bus isn't doing anything else. >And certainly the PC BIOS blocks on read. Woof. No modern operating system uses the BIOS to read disks. BIOS calls rarely work in protected mode. In general, the OS will call a device driver which does the transfers. In fact, this point is moot for most operating systems, as programs don't actually read and write data. In most cases, the program calls the OS which, in the case of a write, writes the data into cache, where it waits to be written to disk. I'd be suprised if even something as horribly twisted as Windows blocks on write. Blocking on reads is more common, mainly because it keeps programs linear. Most operating systems allow non-blocking reads as an option, however. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 23:38:48 GMT Organization: University of Illinois at Urbana-Champaign Lines: 40 Message-ID: <74hou8$n7q$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hlnu$1mq@srvr1.engin.umich.edu> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!news.idt.net!news-a.ais.net!ais.net!vixen.cso.uiuc.edu!tsmurphy In article <74hlnu$1mq@srvr1.engin.umich.edu>, Sergej Roytman wrote: >In article <74hkig$oaq$1@news.wlg.netlink.net.nz>, >Don Stokes wrote: >>In article <1998Dec7.174534.26644@lorelei.approve.se>, >>Goran Larsson wrote: >> Thus, with >>VMS you can mumble: >> >> $ command [...]*.c >> >>and it'll deal to all .c files in the directory and below, regardless of >>how many files are involved, while under unix you end up with: >> >> % command `find . -name='*.c'` > >Huh? > > % command *.c */*.c > >works fine for me. Huh? */*.c will only find files in -immediate- subdirectories. With zsh you can do **/*.c but then you have to use something non-standard (i.e. not supported by your computing support team) And, of course, both of these are subject to the "too many arguments" which was the problem with them in the first place, so you have to resort to xargs hackery (or as the other followup mentioned, creating a file list). Argh!!! I can't believe that VMS is the only mainstream OS which can do right... What's ever sadder is so many people praise Unix's totally broken way of recursing wildcards as elegant (I'd sure like to see -that- explained). Too bad I can't convince my employer to use VMS instead of all this silly Unix and Windows nonsense ;-) -- Terry ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 7 Dec 1998 23:43:30 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 33 Message-ID: <74hp72$lk0$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74hkig$oaq$1@news.wlg.netlink.net.nz>, Don Stokes wrote: > Thus, with >VMS you can mumble: > > $ command [...]*.c > >and it'll deal to all .c files in the directory and below, regardless of >how many files are involved, while under unix you end up with: > > % command `find . -name='*.c'` > >which promptly complains that the argument list is too long and leaves >you farting around with the incredible bogosity that is xargs. Actually, that will execute "command" with no arguments, because you've used the find command incorrectly. The proper method for doing what you describe above is: % find . -name "*.c" -exec command {} \; In the VMS version you haven't told it which file versions to execute the command on. I can't remember whether it would parse "[...]*.c" as "[...]*.c;*" or not. It's been too many years since I used VMS. Can anyone really say they like VMS's "not really a hierarchical file system" structure? Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: viro@vernier.math.psu.edu (Alexander Viro) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 05:08:53 -0500 Organization: -ENOENT Lines: 14 Message-ID: <74itrl$idn@vernier.math.psu.edu> References: <73ela0$51h@hermes.acs.unt.edu> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> NNTP-Posting-Host: vernier.math.psu.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!diablo.theplanet.net!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news-peer.gip.net!news-penn.gip.net!news.gsl.net!gip.net!uwm.edu!news.cse.psu.edu!news3.cac.psu.edu!not-for-mail In article <74glh7$j1n$1@strato.ultra.net>, wrote: >Thus you leave the user or program with the understanding that the >data has been written in iron. Now the next action by the same >user (or, more likely, another user) is to read back the newly >written data. How do you ensure data integrity? Umm? Read back from cache and don't discard a dirty page until you've got a successful completion of IO? Cache belongs to kernel and *which* user's process initiate a request means nothing. It sits between the device and filesystem code. -- "You're one of those condescending Unix computer users!" "Here's a nickel, kid. Get yourself a better computer" - Dilbert. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 09:07:13 +0000 Organization: Tanglewood Message-ID: <366CEC41.A2C9167A@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913108584 nnrp-06:10902 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net!xara.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Richard M. Alderson III wrote: > Is globbing done *in the application*, or is it done by the operating system > via a (set of) system call(s) as in Tops-20? I realize the 32-bit weenies had > a serious case of NIH, but did they really get that wrong, too? To clarify, the wildcard version of the file name is passed to the application, which then makes system calls to obtain a list of files, which match the wildcard. However this happens under control of the application, which can modify the process, and, because it is aware that globbing is going on, can construct related file names. The advantage to this is that the application can control the process, and that there is no limit to the number of files which can be handled. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### Newsgroups: alt.folklore.computers From: Derry Hamilton Subject: Re: Is VMS taboo, or what? In-Reply-To: X-Nntp-Posting-Host: tegan.tardis.ed.ac.uk Content-Type: TEXT/PLAIN; charset=US-ASCII Message-ID: Sender: cnews@dcs.ed.ac.uk (UseNet News Admin) Organization: Department of Computer Science, Edinburgh University References: <73ela0$51h@hermes.acs.unt.edu> <36691be2.2492303@news.uunet.be><366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu><749l84$7vd$1@agate.berkeley.edu> <74afad$nk8@news1.newsguy.com> Mime-Version: 1.0 Date: Tue, 8 Dec 1998 10:17:33 GMT Approved: Me. Lines: 28 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net!xara.net!server6.netnews.ja.net!newsfeed.ed.ac.uk!dcs.ed.ac.uk!tegan.tardis.ed.ac.uk!rasilon On Mon, 7 Dec 1998, Richard M. Alderson III wrote: > Tops-20, and its TENEX predecessor, had swappable and resident portions of the > monitor. One of the folkloric aspects of the first version of TENEX is the > story that the pager code was not originally declared to be resident, leading > to the console error message > > ?PAGE FAULT IN PAGE FAULT HANDLER > > (or similar wording--I was *not* there, ok?) > > I think the Tops-10 monitor was all resident code--but there are those who > will > correct me if I'm wrong, you know. I however was there when NT did that to me. Batch job that requires very little kernel use and is slightly bigger than available memory and is the only thing besides the OS running, works every time.(Image rendering if you must know) Derry Hamilton rasilon@tardis.ed.ac.uk /******************************************************************** * I think your cats need tuning - according to a couple of quick * * measurements on a recently calibrated reference cat, the dominant * * frequency of a correctly adjusted cat should be 12Hz +/-20%. * * ===Lionel Lauer on a.s.r=== * *********************************************************************/ ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec8.101907.3786@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> Lines: 46 NNTP-Posting-Host: 130.244.105.87 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 913112908 130.244.105.87 (Tue, 08 Dec 1998 11:28:28 MET DST) NNTP-Posting-Date: Tue, 08 Dec 1998 11:28:28 MET DST X-Sender: s-64812@dialup105-6-7.swipnet.se Date: Tue, 8 Dec 1998 10:19:07 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article , Neil Franklin wrote: > > Goran Larsson (hoh@approve.se.NO_JUNK_EMAIL) wrote: > > : If the globbing is done in the application then implementing > > : file name completion will be difficult. > > Now why would completing an filename be difficult? If I have typed > "ver" and there is only one file "very_long_named_file" beginning > with "ver" the shell will expand. What has globbing (*,?) got to do with > this? If "VMS scores over" Unix by having the globbing in the application, thus allowing the application to give different meaning to glob patterns, then how is the filename completion code in the "shell" supposed to know how to do a correct filename completion? The simple completion tasks like completing a single filename is trivial, but what should be done if the filename completion is attempted on a partial filename containing glob patterns? > rm * .o can not be caught and ask "do you really want to delete all? It can. Many Unix shells traps this. $ rm * .o zsh: sure you want to delete all the files in /any/path [yn]? > rm "file name with spaces in it *" fails to expand > rm 'file name with spaces in it *' expands and fails, no "file" "name" ... No. Stuff inside single does not expand. The correct way is to write $ rm "file name with spaces in it "* Btw, how can you have spaces in VMS filenames today? > Shell globbing is IMHO the most losing feature in Unix. That is your personal opinion. I Would say quite a number of Unix users doesn't agree with you here. As a former VMS user/programmer/ sysadm (between 1983 and 1987) I can say that VMS has its bright ideas, but having the globbing in the application is not one of them. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec8.103031.4126@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <366C220D.2710F87B@tnglwood.demon.co.uk> Lines: 27 NNTP-Posting-Host: 130.244.205.6 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 913113510 130.244.205.6 (Tue, 08 Dec 1998 11:38:30 MET DST) NNTP-Posting-Date: Tue, 08 Dec 1998 11:38:30 MET DST X-Sender: unknown@dialup205-1-6.swipnet.se Date: Tue, 8 Dec 1998 10:30:31 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article <366C220D.2710F87B@tnglwood.demon.co.uk>, Robert Billing wrote: > Goran Larsson wrote: > > related to globbing do you have with Copying a directory tree? > Try to copy all the *.new files in a tree to *.old in another tree with > a different root but the same topology. cd /a/tree for file in **/*.new; do cp $file /another/tree/`basename $file`.old done > Personally I prefer VMS to unix, and there are a lot of things wrong > with unix, but none of them are fundamental. Replacing the *entire* > shell and command set with something DCL-like would not break unix, and > done carefully it could co-exist with the existing system. It is not enough to create a DCL shell, you also have to create a lot of new tools that fit in with the DCL way of doing things. For example, you need something that acts like COPY instead of cp. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 98 10:33:42 GMT Organization: UltraNet Communications, Inc. Lines: 27 Message-ID: <74j0l4$fc6$1@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> NNTP-Posting-Host: d3.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 8 Dec 1998 10:56:36 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!feed1.news.rcn.net!rcn!news.ultranet.com!d3 In article , alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >In article <366B7CFA.2A976248@tnglwood.demon.co.uk> Robert Billing > writes: > >>Seriously though, the one point on which I think VMS scores over Linux is >>that globbing is done in the application, not the shell, so that you can >>do "related filename" tricks, such as copying a directory tree with the >>copy command. > >Is globbing done *in the application*, or is it done by the >operating system via a (set of) system call(s) as in Tops-20? Rich? Have people changed the meaning of glob? Are they really talking about wildcarding as we in the -10 world knew and loved it? >I realize the 32-bit weenies had >a serious case of NIH, but did they really get that wrong, too? It wouldn't surprise me. There were times in the beginning to go out of their way just to be different from us. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 98 10:42:41 GMT Organization: UltraNet Communications, Inc. Lines: 35 Message-ID: <74j15v$fc6$2@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74itrl$idn@vernier.math.psu.edu> NNTP-Posting-Host: d3.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 8 Dec 1998 11:05:35 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntp.abs.net!feed1.news.rcn.net!rcn!news.ultranet.com!d3 In article <74itrl$idn@vernier.math.psu.edu>, viro@vernier.math.psu.edu (Alexander Viro) wrote: >In article <74glh7$j1n$1@strato.ultra.net>, wrote: >>Thus you leave the user or program with the understanding that the >>data has been written in iron. Now the next action by the same >>user (or, more likely, another user) is to read back the newly >>written data. How do you ensure data integrity? > >Umm? Read back from cache and don't discard a dirty page until you've >got a successful completion of IO? Cache belongs to kernel and *which* >user's process initiate a request means nothing. It sits between the >device and filesystem code. > This is just fine in a single-user system. But what do you do if you've got 500 users updating the same data base where the data base consists of many, many little files (I know this would be a very bad application design but I'm trying to take an extreme to get you thinking about good OS design). There's no right or wrong answer. The approach of the implementation really depends on the phiolosphy of the operating system design. What are strengths in one is a weakness in another and visa versa. Another thing to consider is efficiency. In one OS, your suggestion is great; in another OS the suggestion is a terrible w.r.t. performance. All of these aspects of implementing an OS need to be considered. The decisions made start to mold the philosophy of the OS. Some outlast drastic hardware evolutions and some don't. /BAH /BAH Subtract a hundred and four for e-mail. ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec8.105505.4519@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> Lines: 72 NNTP-Posting-Host: 130.244.241.82 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 913114710 130.244.241.82 (Tue, 08 Dec 1998 11:58:30 MET DST) NNTP-Posting-Date: Tue, 08 Dec 1998 11:58:30 MET DST X-Sender: unknown@dialup241-2-18.swipnet.se Date: Tue, 8 Dec 1998 10:55:05 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article <74hkig$oaq$1@news.wlg.netlink.net.nz>, Don Stokes wrote: > In article <1998Dec7.174534.26644@lorelei.approve.se>, > Goran Larsson wrote: > > >If the globbing is done in the application then implementing > >file name completion will be difficult. What kind of difficulties > > I don't know what you mean by file name completion being difficult. Having > implemented DCL filename completion, believe me, the completion bit was > easy. (The bit to get the completion routine actually called by DCL was > a *wee* bit trickier... 8-) How did you tell the completion routine how the application wanted the filename completion to be done? If you have the files TEST.TXT and TEST.EXE, then how did you handle this? TYPE TE RUN TE also TYPE T*;-1 > Shell globbing isn't that great. Shell globbing means all you need > to process a globbed list is: > > for(i = 1; i < argv; i++) process(argv[i]); > > > by contrast, using LIB$FIND_FILE: > > for(i = 1; i < argv; i++) while(file = find_file(argv[i])) process(file); The Unix way is consistent, even if it makes no sense for some programs. You know it is there and if you know how it works there will be no surprices. The DCL way leaves the users at the mercy of the programmer, you never know if the application handles it the way you expect. In the end it is just a user preference. If all users wanted it the DCL way, then by now we would have a number of Unix shells doing it that way, or? > has to be sucked into memory by the shell, and handed bodily to the > subprocess. And the shell doesn't have recursive searches. Thus, with > VMS you can mumble: > > $ command [...]*.c $ command **/*.c All files matching *.c $ command **/*.c(u:hoh:) ...and owned by user 'hoh' $ command **/*.c(m-2) ...and modified within the last 2 days Note that in all these examples I can press TAB instead of RETURN and have the pattern replaced by whatever it expands to, e.g. $ ls **/c*(m-2) $ ls config.cache config.h config.log config.status src/crc.h The shell (zsh) replaced the command line with the expanded command line for me. I have no idea how this could be implemented in DCL. > % command `find . -name='*.c'` Make that $ command `find . -name='*.c' -print` to be portable as POSIX requires that -print be given if intended. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 11:04:26 GMT Organization: Daedalus Consulting Lines: 33 Message-ID: <74j13q$6c3$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Howard S Shubs wrote: >korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >>Can anyone really say they like VMS's "not really a hierarchical file system" >>structure? > >How is it not hierarchical? I'd assume he's talking about the fact that under VMS, the device, directory, file name, type and version are all distinct syntactic elements, as opposed to unix where there is a path name, and the elements in the path can be devices, directories, file names, symbolic links, mount points or various other things, all occupying one hierarchical name space. Well, VMS doesn't do things that way. Cope. What VMS does do is use logical names extensively. VMSsy (VMeSsy?) applications get at files etc as logicalname:file or even just the logical name alone (and there's magic in the file parsing stuff that makes this quite flexible). Logical names exist outside the filesystem namespaces; they can be system, group, job (process group in unix parlance) or process specific, so you can run two incarnations of the same application simultaneously with the same hard-coded filenames but acting on different files. This is a rather difficult model to move to unix -- the closest you can get is to create environment variables and have your application translate them before opening the file, and environment variables are notoriously difficult to apply systemwide. VMS logical names are only too happy to be changed systemwide on the fly. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: deke.spamblock@generous.net Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 12:50:49 GMT Organization: GenerousCity is a virtue - find romance at http://generous.net Lines: 23 Message-ID: <366d1f6c.9761743@news.bright.net> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: paul-cas2-cs-29.dial.bright.net X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!news.indiana.edu!NewsNG.Chicago.Qual.Net!news.bright.net!not-for-mail On Mon, 07 Dec 1998 21:21:52 -0600, hshubs@mindspring.com (Howard S Shubs) wrote: >>Hmmm, on my unix-like system when I type in "help" I get... >Does straight UNIX do that? And, does HELP work on its own, w/o referring >to MAN? The UNIX command set is waaaaay arbitrary and counter-intuitive. "Straight UNIX" doesn't do anything. It doesn't interface with the user at all. You need a shell program to process user input. If you don't like the B shell or the C shell, or the Korn shell or the BASH shell, write your own shell! After all, it's not like writing a shell is rocket science.... deke ------------------------ Let love find you! http://generous.net A list for flirting GenerousSingles-subscribe@onelist.com Over The Hill Gang GenerousSinglesOTHG-subscribe@onelist.com Personal Ads list GenerousProfiles-subscribe@onelist.com ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 16:22:21 GMT Organization: University of Illinois at Urbana-Champaign Lines: 9 Message-ID: <74jjnt$i01$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <366d1f6c.9761743@news.bright.net>, wrote: >"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. >You need a shell program to process user input. Please point me to a system which is branded UNIX and does not contain a Bourne shell. If you can't please take back your statement. -- Terry ###### From: TheCentralScrutinizer.100@pobox.com () Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 16:36:07 GMT Organization: http://extra.newsguy.com Lines: 13 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> Reply-To: TheCentralScrutinizer.100@pobox.com NNTP-Posting-Host: edison.chisp.net X-Newsreader: slrn (0.9.3.2 UNIX) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!TheCentralScrutinizer.100 On 8 Dec 1998 16:22:21 GMT, TERENCE MURPHY wrote: >In article <366d1f6c.9761743@news.bright.net>, > wrote: >>"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. >>You need a shell program to process user input. > >Please point me to a system which is branded UNIX and does not contain >a Bourne shell. If you can't please take back your statement. When I first used unix, we didn't even have hardware to run it on. We did all our computations on paper. And we liked it! ###### From: atbowler@thinkage.on.ca (Alan Bowler) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 17:24:05 GMT Organization: Thinkage Ltd. Lines: 13 Message-ID: <74jnbl$kmc$1@nntp3.uunet.ca> References: <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> NNTP-Posting-Host: 192.102.11.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!newsfeed1.earthlink.net!uunet!in2.uu.net!news.uunet.ca!atbowler In article <366858fd.4481731@news.vip.net> genew@vip.net writes: > > I thought that was how it would be. What I was really worried >about was how it would handle when a swap was needed, especially a >swap IN. So, in the example given, the user was lucky that no >swapping was required? A job unlucky enough to be swapped out to a hung drive will wait for the drive to become available. If that drive fails completely, the jobs on that disks would have to be shot.I don't know what VMS does, but there are systems that spread their backing store (swap and/or paging) over multiple disks so the failure of a single drive will often only affect a few processes. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 18:00:28 +0000 Organization: Tanglewood Message-ID: <366D693C.DF21812@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913142828 nnrp-10:17335 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 16 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail TERENCE MURPHY wrote: > Please point me to a system which is branded UNIX and does not contain > a Bourne shell. If you can't please take back your statement. Ass you have to do is install, say, zsh, and remove the Bourne shell, and there you are. It is completely possible to construct a Linux system with no shells at all (if not very useful) simply by nobbling the startup mechanics to run a single application rather than init or a shell. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 18:05:19 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 49 Message-ID: <74jpov$4iu$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74j13q$6c3$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >Howard S Shubs wrote: >>korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >>>Can anyone really say they like VMS's "not really a hierarchical file system" >>>structure? >> >>How is it not hierarchical? > >I'd assume he's talking about the fact that under VMS, the device, >directory, file name, type and version are all distinct syntactic >elements, as opposed to unix where there is a path name, and the >elements in the path can be devices, directories, file names, symbolic >links, mount points or various other things, all occupying one >hierarchical name space. The VMS file system has always felt to me like it is something cobbled onto a flat file system. I always got this feeling that if I got a Micro-11 to mount a VMS disk it would show all the files just sitting there in the root directory. For fun, copy your favorite executable to "NEW.DIR" >Well, VMS doesn't do things that way. Cope. My method of coping was to cease using VMS. Is there anyone out there who didn't alias "SET DEFAULT" to something else? "CD" maybe? >What VMS does do is use logical names extensively. That's only because it's too damn messy to type "BKYAST::SYS$DUA0:[USERS.KORPELA.STATS]". It's a bug fix, not a feature. >This is a rather difficult model to move to unix -- the closest you can >get is to create environment variables and have your application >translate them before opening the file, and environment variables are >notoriously difficult to apply systemwide. VMS logical names are only >too happy to be changed systemwide on the fly. Well, it wouldn't be that difficult to create the equivalent of logical names in a unix shell. It's just not really necessary. Why create something like that when a symbolic link will work. Why should accessing a disk mounted over the network look any different than accessing a local disk? Why even continue the concept of separate disks? It makes moving things from one disk to another difficult. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <366f6b01.9489096@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 18 Date: Tue, 08 Dec 1998 18:09:47 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 913140580 129.85.24.56 (Tue, 08 Dec 1998 13:09:40 EDT) NNTP-Posting-Date: Tue, 08 Dec 1998 13:09:40 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On Mon, 07 Dec 1998 21:28:34 -0600, hshubs@mindspring.com (Howard S Shubs) wrote: >In article <74hp72$lk0$1@agate.berkeley.edu>, >korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: [ snip ] >>Can anyone really say they like VMS's "not really a hierarchical file system" >>structure? > >How is it not hierarchical? And while we are at it, how in "hierarchical" Unix file system one can determine the (used) size of "/" partition (excluding other file systems hanging off /) ? [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> X-Newsreader: trn 4.0-test67 (15 July 1998) From: cmikk@tig.oss.uswest.net (Chris Mikkelson) NNTP-Posting-Host: 204.147.85.37 Message-ID: <366d8031.0@news2.uswest.net> Date: 8 Dec 1998 13:38:25 +0600 X-Trace: 8 Dec 1998 13:38:25 +0600, 204.147.85.37 Lines: 12 X-Report: Report abuse to abuse@uswest.net. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!logbridge.uoregon.edu!enews.sgi.com!su-news-feed4.bbnplanet.com!news.gtei.net!newsfeed1.uswest.net!news1.uswest.net!not-for-mail In article <366f6b01.9489096@Rockyd>, > >And while we are at it, how in "hierarchical" Unix file system one can determine Do you mean: (cmikk@tig) ~ $ df / Filesystem 512-blocks Used Avail Capacity Mounted on /dev/sd0s1a 77982 30014 41730 42% / (cmikk@tig) ~ $ -Chris ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <366f6b01.9489096@Rockyd> <366d8031.0@news2.uswest.net> <367591ad.19390965@Rockyd> X-Newsreader: trn 4.0-test67 (15 July 1998) From: cmikk@tig.oss.uswest.net (Chris Mikkelson) NNTP-Posting-Host: 204.147.85.37 Message-ID: <366dabe5.0@news3.uswest.net> Date: 8 Dec 1998 16:44:53 +0600 X-Trace: 8 Dec 1998 16:44:53 +0600, 204.147.85.37 Lines: 41 X-Report: Report abuse to abuse@uswest.net. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!su-news-hub1.bbnplanet.com!su-news-feed4.bbnplanet.com!news.gtei.net!newsfeed1.uswest.net!news1.uswest.net!not-for-mail In article <367591ad.19390965@Rockyd>, Alexandre Pechtchanski wrote: >On 8 Dec 1998 13:38:25 +0600, cmikk@tig.oss.uswest.net (Chris Mikkelson) wrote: >>Do you mean: >> >>(cmikk@tig) ~ $ df / >>Filesystem 512-blocks Used Avail Capacity Mounted on >>/dev/sd0s1a 77982 30014 41730 42% / >>(cmikk@tig) ~ $ > >No, I mean if >clinfo> df / >Filesystem Total kbytes kbytes % >node kbytes used free used Mounted on >/dev/rz2a 19743 19337 0 109% / > ^^^^^ ^^^^ The Unix filesystem, IIRC, reserves some free space (i.e. temporary space for reorganizing files) that only root can use for files. This accounts for >100% usage. >the numbers don't tally as above (I did pain-stakingly added all file sizes in / >and came up with different figure), how can I persuade Unix to do sums for me >(on VMS I would simply do DIR [...]/TOTAL) Since each disk/FS block can only contain data from one file, a file will *use* more disk space than its size would reflect. Thus, df reports the real used and free space on the disk, rather than just a total of the file sizes. Can you do that on VMS? Not being snide, just asking. The equivalent of DIR /TOTAL (which doesn't list the entries, right?) would be du (disk usage), which will cross mount points. I can say this is equivalent to the VMS version, assuming that the VMS "DIR /TOTAL" would cross mount points if VMS had mount points. Oh, and why did you "painstakingly" add all the file sizes when you could use awk? -Chris ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 19:36:50 GMT Organization: University of Illinois at Urbana-Champaign Lines: 34 Message-ID: <74jv4i$t2a$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <74jpov$4iu$1@agate.berkeley.edu>, Eric J. Korpela wrote: >My method of coping was to cease using VMS. Is there anyone out there >who didn't alias "SET DEFAULT" to something else? "CD" maybe? ^^^^^^ I don't (note the tense) alias "SET DEFAULT" and on Unix systems I alias "setdef" to be cd. Unfortunately, Unix isn't as customizable as VMS so I can't set up a multi-word alias. >That's only because it's too damn messy to type >"BKYAST::SYS$DUA0:[USERS.KORPELA.STATS]". It's a bug fix, not a feature. It's not a bug fix, it does what symlinks were supposed to do (but miserably failed at). Standard directories have logical names. This is so you don't need to know where stuff is. For example, all of the include files for DECwindows are in a standard directory. Can Unix say the same? The supposedly standard /usr/include/X11, for example, doesn't work on my Solaris and HP-UX accounts. So the only way to find stuff on Unix is to use find which takes about 8 hours to complete. >Well, it wouldn't be that difficult to create the equivalent of logical >names in a unix shell. It would be easy to do in the shell but it would require a complete re-design and re-write of every program. >Why should accessing a disk mounted over the network look any different >than accessing a local disk? Why even continue the concept of separate >disks? It makes moving things from one disk to another difficult. You are entirely missing the point of logicals. -- Terry ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 20:33:29 -0600 Organization: MindSpring Enterprises Lines: 25 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: d1.56.d1.de X-Server-Date: 9 Dec 1998 02:33:30 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!howland.erols.net!news-peer1.sprintlink.net!news.sprintlink.net!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <74j13q$6c3$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz (Don Stokes) wrote: >Howard S Shubs wrote: >>korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >>>Can anyone really say they like VMS's "not really a hierarchical file system" >>>structure? >> >>How is it not hierarchical? > >I'd assume he's talking about the fact that under VMS, the device, >directory, file name, type and version are all distinct syntactic >elements, as opposed to unix where there is a path name, and the >elements in the path can be devices, directories, file names, symbolic >links, mount points or various other things, all occupying one >hierarchical name space. If that's what he means, I agree with your conclusion: deal with it. It's good for keeping things separate, as opposed to under UNIX where the division between a disk and a directory is unclear. "Well, this directory contains files x y and z unless -that- partition is mounted, in which case it contains files a b and c. So, how do you get at x y and z to append them to a b and c? Gotta use a scratch area. Yeah, that's smooth." -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 20:43:58 -0600 Organization: MindSpring Enterprises Lines: 62 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: d1.56.d1.de X-Server-Date: 9 Dec 1998 02:44:01 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74jpov$4iu$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >The VMS file system has always felt to me like it is something cobbled onto >a flat file system. I always got this feeling that if I got a Micro-11 >to mount a VMS disk it would show all the files just sitting there in >the root directory. Strange. Probably due to a lack of experience with VMS vs some experience with RSX-11. >For fun, copy your favorite executable to "NEW.DIR" It'll still run, normal executables don't -have- to have the .EXE extension, but doing a DIRECTORY on it will not. >My method of coping was to cease using VMS. Is there anyone out there >who didn't alias "SET DEFAULT" to something else? "CD" maybe? Yeah, we use set def, or SET DEF, or whatever, as the mood and the CAPS LOCK key take us. Can you do that? Without using "alias"? >That's only because it's too damn messy to type >"BKYAST::SYS$DUA0:[USERS.KORPELA.STATS]". It's a bug fix, not a feature. I'd kick a user who did it this way. Probably by moving the directory from one disk to another, or by renaming the machine. Done it before, after warning that stuff was going to move. for some reason or other. Logical names are, in part, a way for system managers to alias things in order to organize the system better. As a UNIX-head, you're familiar with the concept of the "~" directory. Which is better, using that notation or using the full path? >Well, it wouldn't be that difficult to create the equivalent of logical >names in a unix shell. It's just not really necessary. LOL! Without a real understanding of logical names other than as shortcuts for files and directories, you're trying to draw conclusions about them?? >Why should accessing a disk >mounted over the network look any different than accessing a local disk? >Why even continue the concept of separate disks? It makes moving things >from one disk to another difficult. Security, for one thing. For another, perhaps it's more efficient to make things clearer for the users. Group X owns all the space on disks 1..20, so they -know- they shouldn't expect access to other disks, for instance. If disk 1 is on a machine across the planet, and disk 2 is connected to their box, it's generally useful to know that so you don't require a lot of swap-like accesses to disk 1. These kinds of rwars are fun! -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 20:51:11 -0600 Organization: MindSpring Enterprises Lines: 23 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> NNTP-Posting-Host: d1.56.d1.de X-Server-Date: 9 Dec 1998 02:51:12 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsfeed.ecrc.net!207.69.200.14.MISMATCH!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <74k6sf$dr5$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >I think you miss the point entirely. Like wildcard expansion, logical >names would be expanded by the shell, not by the programs. You mean, like DCL symbols are? >In fact every >unix system I've used has shells that support the logical name "~" which >in VMS speak would be "HOME:". You mean SYS$LOGIN. >The programs don't need to know how to >expand "~" because they don't see it. LOL!!!! I just wrote a bit of code in a program to handle "~" because the users can type it into the program as a shorthand. Maybe they don't -need- to know, but it can be useful. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 08 Dec 1998 20:54:55 -0600 Organization: MindSpring Enterprises Lines: 13 Message-ID: References: <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> <74jnbl$kmc$1@nntp3.uunet.ca> NNTP-Posting-Host: d1.56.d1.de X-Server-Date: 9 Dec 1998 02:54:56 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!nntp.abs.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74jnbl$kmc$1@nntp3.uunet.ca>, atbowler@thinkage.on.ca (Alan Bowler) wrote: >the jobs on that disks would have to be shot.I don't know what VMS does, >but there are systems that spread their backing store (swap and/or >paging) over multiple disks so the failure of a single drive will often >only affect a few processes. You can have multiple page files, which can be on different disks. Still, a process which has space paged out on some disk when that disk dies, is in trouble. -- Howard S Shubs The Denim Adept ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <367591ad.19390965@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> <366d8031.0@news2.uswest.net> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 25 Date: Tue, 08 Dec 1998 21:00:50 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 913150844 129.85.24.56 (Tue, 08 Dec 1998 16:00:44 EDT) NNTP-Posting-Date: Tue, 08 Dec 1998 16:00:44 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On 8 Dec 1998 13:38:25 +0600, cmikk@tig.oss.uswest.net (Chris Mikkelson) wrote: >In article <366f6b01.9489096@Rockyd>, >> >>And while we are at it, how in "hierarchical" Unix file system one can determine > >Do you mean: > >(cmikk@tig) ~ $ df / >Filesystem 512-blocks Used Avail Capacity Mounted on >/dev/sd0s1a 77982 30014 41730 42% / >(cmikk@tig) ~ $ No, I mean if clinfo> df / Filesystem Total kbytes kbytes % node kbytes used free used Mounted on /dev/rz2a 19743 19337 0 109% / ^^^^^ ^^^^ the numbers don't tally as above (I did pain-stakingly added all file sizes in / and came up with different figure), how can I persuade Unix to do sums for me (on VMS I would simply do DIR [...]/TOTAL) [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 21:30:58 GMT Organization: Daedalus Consulting Lines: 19 Message-ID: <74k5qi$h9t$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article , Richard M. Alderson III wrote: >Is globbing done *in the application*, or is it done by the operating system >via a (set of) system call(s) as in Tops-20? I realize the 32-bit weenies had >a serious case of NIH, but did they really get that wrong, too? It's in RMS (Record Management Services). RMS calls are handled as system services rather than linked in library calls (as in RMS-11), but sit on top of the basic file system primitives built into the ACP/XQPs. So yeah, it's pretty well buried in the OS. LIB$FIND_FILE which is what mere mortals use to do globbing is a library call and just front-ends the RMS stuff. The RMS calls are actually pretty easy to use if you're used to RMS's ways of doing things, but setting up FAB & NAM blocks tends to be a bit annoying in high level languages; MACRO makes it pretty easy though. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 21:49:03 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 21 Message-ID: <74k6sf$dr5$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74jv4i$t2a$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: > >>Well, it wouldn't be that difficult to create the equivalent of logical >>names in a unix shell. > >It would be easy to do in the shell but it would require a complete >re-design and re-write of every program. I think you miss the point entirely. Like wildcard expansion, logical names would be expanded by the shell, not by the programs. In fact every unix system I've used has shells that support the logical name "~" which in VMS speak would be "HOME:". The programs don't need to know how to expand "~" because they don't see it. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 22:10:40 GMT Organization: Daedalus Consulting Lines: 68 Message-ID: <74k850$996$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74jpov$4iu$1@agate.berkeley.edu>, Eric J. Korpela wrote: >The VMS file system has always felt to me like it is something cobbled onto >a flat file system. I always got this feeling that if I got a Micro-11 >to mount a VMS disk it would show all the files just sitting there in The directory/file relationship is not significantly different to unix. Files are described by file headers (roughly analogous to unix inodes), and directories contain file names and file IDs. The ODS2 filesystem is a direct descendant of RSX Files-11 (aka ODS1); VMS will happily create and/or mount an RSX filesystems, but not vice-versa. ODS1 can handle directories, but RSX only knows about directories within the root directory (and until fairly recently only numeric directories), but the filesystem format never precluded subdirectories. If you create an ODS1 filesystem undeer VMS, create a [X.Y] directory and mount it under RSX, you'll see a Y.DIR in [X], but won't be able to use it. RSTS and RT-11 have different filesystems. >Well, it wouldn't be that difficult to create the equivalent of logical >names in a unix shell. It's just not really necessary. Why create something >like that when a symbolic link will work. Why should accessing a disk Logical name translation lives in the filesystem, and apply to any file opened by an application, not just those handed to it by the shell. I already pointed out why logical names are different and in some ways more useful than symlinks; if you're going to turn this into an OS advocacy argument and willfully ignore what's already been said then take it to the appropriate advocacy group. That's not to say I don't think symlinks are a good idea -- I use 'em all the time with unix. I use logical names with VMS. Symlinks are a unix way of doing things, logicals are VMS's. They both have their merits and demerits. (Actually, ODS2 specifies symlinks in the directory structure, but nobody ever implemented them.) >mounted over the network look any different than accessing a local disk? DECNET filenames don't require remote mounting to happen at all -- the nodes are on the network, and by giving a filename of node::filename, RMS will happily forward the request to the remote node. No setup required. If you remotely mount a disk (via NFS or DFS), you'll just get a different device name. >Why even continue the concept of separate disks? It makes moving things >from one disk to another difficult. cp /mnt-pt/dir/file /mnt-pt2/newdir/newfile isn't really any different from: copy disk1:[dir]file disk2:[newdir]file If you want shorthand, assign appropriate logical names -- you can assign them on the fly. Unix doesn't have a hierarchical filesystem as such (at least no more than VMS does, and both VMS and unix have the same limitations in what you can and can't do between filesystems ), but it does have a single hierarchical namespace. VMS has multiple, somewhat less hierarchical namespaces -- the device names, per-process, per-job, per-group and system logical name tables, and filesystem names, several of which are consulted in looking for files. The unix approach does have a naive charm, but it doesn't allow for per-process or per-session names to be assigned. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: dicks@math.ohio-state.NO.SPAM.edu (Ethan Dicks) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 22:57:15 GMT Organization: Department of Mathematics, The Ohio State University Lines: 25 Message-ID: <74kasb$47f$1@mathserv.mps.ohio-state.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> NNTP-Posting-Host: math.mps.ohio-state.edu X-Trace: mathserv.mps.ohio-state.edu 913157835 4335 128.146.111.30 (8 Dec 1998 22:57:15 GMT) X-Complaints-To: usenet@math.ohio-state.edu NNTP-Posting-Date: 8 Dec 1998 22:57:15 GMT X-Newsreader: trn 4.0-test68 (29 August 1998) Originator: dicks@math.ohio-state.edu (Ethan Dicks) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!math.ohio-state.edu!not-for-mail In article <366f6b01.9489096@Rockyd>, Alexandre Pechtchanski wrote: > >And while we are at it, how in "hierarchical" Unix file system one can >determine the (used) size of "/" partition (excluding other file systems >hanging off) ? > >[ When replying, remove *'s from address ] >Alexandre Pechtchanski, Systems Manager, RUH, NY Under Solaris, du -d will avoid crossing filesystem boundaries. I typically use "du -skd" to give me a report of the total size (not intermediate directories), in K bytes, only on this filesystem. I also like "su -s *" which tells me where all the space has gone underneath my current directory. -ethan (Who used VMS in a former life and liked it) -- Ethan Dicks http://www.infinet.com/~erd/ (dicks) at (math) . (ohio-state) . (edu) sellto: postmaster@[127.0.0.1] harvestbot fodder: president@whitehouse.gov fccinfo@fcc.gov root@[127.0.0.1] ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 8 Dec 1998 23:11:52 +0100 Organization: [Posted via] Interactive Networx Lines: 49 Message-ID: <1djq6kf.1opse7p8w8fuoN@n32-122.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: n32-122.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!newsfeed.ecrc.net!unlisys!news.snafu.de!jnickelsen Tim Shoppa wrote: > On the other hand, VMS people develop an intense dislike for Unix [...] > I had an executable called "test" in my current directory that was > supposed to do a couple simple calculations and display the result. So I > type "test" to run my new program - and nothing comes out. [...] > Of course, eventually I learned my first "school of hard knocks" lesson > from Unix, and the necessity of the "which" command, and how it relates to > the PATH environment variable. But it took us several hours to figure > this out. Sure, unobvious properties of the environment (in the broader sense) can make your life *really* hard. But I don't think VMS's corresponding unobvious properties (the command name space, logicals, DCL symbols) are fundamentally easier to grasp for Unix wee^H^H^Hpeople than Unix's PATH etc. are to grasp for VMS people. I know it took me some time, and I wasn't much thinking in Unix terms then. Of course, the consistency of the VMS (or DCL) command syntax makes lots of things easy that are difficult to grasp under Unix. But PATH, on the other hand, is one of Unix's consistent concepts. Concerning the use of command line qualifiers/options (e. g. "-r") I agree with you completely. This is a *mess* with Unix. My most disfavourite example is the difference between the commands "tcpdump" (BSD and several others) and "snoop" (Solaris). Both commands do basically the same: grab network packets from an interface and display their header (and optionally payload content) in some way. There are some differences in their capabilities, but these are minor; most of what both commands do is the same. As expected, the expressions to select the packets which you are interested in are the same, along the lines of "ip multicast and not broadcast and ( host picasso or host monet )". But: for some reason which completely escapes me, the command options are typically different! Write output to a file is "-o" with snoop, "-w" with tcpdump, read input from a file is "-i" with snoop and -r with tcpdump (or the other way 'round, I can't remember). Select a certain interface is "-i" vs. "-I" etc. etc. Yeah, well. I still like Unix, since I am at home with it, and the alternative is usually Windows NT. -- Juergen Nickelsen ##### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 8 Dec 1998 23:11:55 +0100 Organization: [Posted via] Interactive Networx Lines: 19 Message-ID: <1djq8eg.js1nvtpewd96N@n32-122.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: n32-122.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.ecrc.net!unlisys!news.snafu.de!jnickelsen Don Stokes wrote: > Shell globbing isn't that great. Shell globbing means all you need > to process a globbed list is: > > for(i = 1; i < argv; i++) process(argv[i]); > > > by contrast, using LIB$FIND_FILE: > > for(i = 1; i < argv; i++) while(file = find_file(argv[i])) process(file); > > ("find_file" is a (fictional) wrapper to lib$find_file [...] I wrote a setargv(&argc, &argv) function once using LIB$FIND_FILE. If anyone is interested... -- Juergen Nickelsen ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 8 Dec 1998 23:16:51 GMT Organization: University of Illinois at Urbana-Champaign Lines: 26 Message-ID: <74kc13$chj$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <74k6sf$dr5$1@agate.berkeley.edu>, Eric J. Korpela wrote: >I think you miss the point entirely. Like wildcard expansion, logical >names would be expanded by the shell, not by the programs. In fact every >unix system I've used has shells that support the logical name "~" which >in VMS speak would be "HOME:". The programs don't need to know how to >expand "~" because they don't see it. Well, you predicted the fault of your own argument: logicals need to be expanded in interactive programs, not just the shell. As for ~ only being handled by the shell, that's false of course many interactive programs handle it (I just tried my newsreader, trn, and it does). (Which makes me wonder, is there a standard library routine that I don't know about which does this or is each program hacking its own?) Anyway's what's annoying about this of course is that some programs can do it and some can't (the usual Unix fashion of inconsistency)...just a couple of days ago I created a file with a ~ in it because some program wasn't doing it right (think it was make...) Also, this is the answer to claims that this sort of stuff should be handled in the shell not the program. -- Terry ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: jmfbahciv@aol.com's message of Tue, 08 Dec 98 10:33:42 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <74j0l4$fc6$1@strato.ultra.net> Date: Wed, 9 Dec 1998 00:52:31 GMT Lines: 17 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-a.ais.net!ais.net!ix.netcom.com!netcom!alderson In article <74j0l4$fc6$1@strato.ultra.net> jmfbahciv@aol.com writes: >Rich? Have people changed the meaning of glob? Are they really talking about >wildcarding as we in the -10 world knew and loved it? That's what it's used for, although it's not as efficient as the Tops-20 method --I'm not sure how Tops-10 handled this--since the glob() system call builds a list data structure in memory pre-allocated by the program, where the GTJFN% and GNJFN% JSYi chase things down through the directory data block(s) (sort of like i-nodes for the Unix readers, but not really). It's the only meaning of "glob" I know--I only ever heard it in a Unix context. What else does it refer to? -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 00:59:12 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 36 Message-ID: <74ki10$ln4$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74k850$996$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74k850$996$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >>Why even continue the concept of separate disks? It makes moving things >>from one disk to another difficult. > > cp /mnt-pt/dir/file /mnt-pt2/newdir/newfile > >isn't really any different from: > > copy disk1:[dir]file disk2:[newdir]file I was thinking more on the lines of.... % cd /mnt-pt2/newdir % tar cf - -C /mnt-pt/dir . | tar xf - % /bin/rm -rf /mnt-pt/dir % ln -s /mnt-pt2/newdir /mnt-pt/dir Software installed on disk1:[dir] may not like being moved to disk2:[newdir]. Users may not be able to find it. So the VMS administrator ends up defining a logical name for the directory of every piece of software installed on the system. If you're lucky that is. Many administrators I've worked with didn't even tell the users when new software was installed, much less setup logicals for it. And given that users rarely have read permission in disk root directories under VMS, the only way to find that piece of software is to call the administrator and ask them if it is installed. About the only thing I missed when moving from VMS to UNIX is SHOW PROCESS. Now that the /proc file system is here, I don't even miss that. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 01:47:16 GMT Organization: Micro Focus Inc. Lines: 91 Message-ID: <74kkr4$6t9@news1.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> NNTP-Posting-Host: p-286.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!mww In article <74glh7$j1n$1@strato.ultra.net>, jmfbahciv@aol.com writes: > [somebody's hacked off the atributes and it tweren't me] > > In article <366b39eb.3006571501@news.cyberport.com>, > tangentSPAMCATCHER@cyberport.com (Warren Young) wrote: > >I've only been working with computers seriously since about 1990 (and > >a nerdy little whiz kid since 1985), but it seems to me that you could > >do this with a preemptible, multitasking kernel. No, not a kernel > >that provides multitasking to the user programs, but that can be > >preempted itself. This is probably easiest to do with a > >microkernel-style implementation. I know NT has a semi-micro-kernel > >setup, and I suspect that VMS does, too. > > You must consider the case where your microkernel preempts itself. The AIX kernel, as of v3, is preemptible as well as pageable. And it's not even a microkernel. It's not a trivial design, but it's not prohibitively difficult. Mostly it means putting a lot of fast synchronization mechanisms in at the right places. > >So, when the swapper or pager process decides it needs to move a block > >of memory to the swap file to free up some memory, it simply calls the > >disk I/O routines to do this. They then realize that the volume in > >question is apparently unavailable. Now, since it hasn't been > >formally dismounted, it's probably a temporary error, so the call > >blocks. The pager then sits there, blocking on a disk I/O call until > >it completes. AIX, like some other modern Unix implementations, doesn't have a separate pager process. All disk I/O (except I/O to special partitions that are reserved to special third-party applications, like DBMSes, and largely ignored by the OS) goes through the kernel's virtual memory manager. That means all swap and file I/O is memory-mapped, under the covers ("covers" being the read, write, etc. system calls) if necessary. The VMM, which is a kernel lightweight process, won't block forever on I/O problems because it has a watchdog timer. I haven't looked into how the kernel behaves in these cases, but I imagine its action is based on the severity of the problem (probably transient? probably permanent? affecting user code? a critical kernel process?). > Assuming it is a temporary error (whatever that means...there are > myriads of categories), do you tell someone about the error? And, > if you do think someone should be notified, who? In Unix, that's a configuration issue, if the problem is non-fatal; it's logged via the syslog facility or something similar in most flavors, and you have various options for what "logging" involves. If the kernel panics, that's a different story, but again details are implementation-specific. Typically the kernel dumps core to the save device, if it can, and shuts down or reboots. Many modern machines have some facility for storing some error information in non-volatile memory, too. > >Alternately, that disk I/O call could complete after passing the data > >off to the cache subsystem. The cache subsystem would then wait > >patiently for the disk to become available again. > > Thus you leave the user or program with the understanding that the > data has been written in iron. Now the next action by the same > user (or, more likely, another user) is to read back the newly > written data. How do you ensure data integrity? With the all-I/O-via-VMM architecture, it's not a problem. The OS knows which pages are in RAM and which are not. As you pointed out in another note in this stream, there are performance implications, but actually the AIX architecture performs very well for the situation you describe, given a reasonable amount of RAM. That's partly because it has hardware assist in the form of reverse translation lookaside buffers. The OS does provide a system call to force immediate scheduling of a dirty page commit, for applications that want a little more assurance; but of course many drive/controller combinations these days will take queued requests and pretend they're done before they've actually been committed to physical media. But as you say that's not an ideal architecture for all applications, which is why you have DBMSes that write to raw partitions and don't go through the OS' VMM. Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University Love thine and hamster and live with the hamster. < From the hamster and the sutra > -- roro ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 02:36:44 GMT Organization: Micro Focus Inc. Lines: 148 Message-ID: <74knnt$8c1@news1.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: p-604.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!mww In article , hshubs@mindspring.com (Howard S Shubs) writes: > In article <36690D71.4F611118@trailing-edge.com>, Tim Shoppa > wrote: > > >On the other hand, VMS people develop an intense dislike for Unix, > >finding fatal design flaws in the command interface almost immediately. > What -I- see in UNIX: [rearranged slightly for clarity] > 1) poor command-line design [rant snipped] I've used an awful lot of OSes, and I've never seen one with an intuitive UI. Yes, the names of the commonly used Unix commands are obscure. I didn't find them any more difficult to learn than those of a dozen other platforms. But this is so thoroughly subjective an issue that it's hardly worth mentioning, let alone discussing. > 2) lack of developer support other than the "make" command > > (2) Let me ask you about the UNIX source-level debuggers. So, if you get > a crash, you can use the "core" file. Great. What if you want to find > something being changed inapproprately? A watchpoint takes FOREVER, and > in my experience, usually CRASHES THE DEBUGGER after taking ALL NIGHT to > find what I've been looking for. The debuggers under UNIX make me go back > to pre-debugger technologies of "write value of variable at various points > in the program". Yes, that's advanced. Watchpoints are a hardware issue; they have nothing to do with software. Either there's hardware assist (like watch registers) or there isn't. (If the hardware assist is there, but the software doesn't use it, get a different debugger. They're generally free.) Unix has developer tools all over the place. That's mostly what Unix is - a collection of developer tools (though again that's not a core OS matter). I humbly submit that you don't know what you're talking about here. (That may be because Unix is "hard to learn" or some such, but that's an entirely different matter.) > How much cooperation do you get on UNIX between languages? Okay, I can do > sorta kinda okay (Contortions 'R' Us) using f77 and cc, but if I wanted to > use other language, what new obscurities would I need to discover? An implementation issue. Has nothing to do with Unix. If you stick to one company's Unix implementation, there frequently isn't a problem. There's nothing involved in calling between Fortran and C on AIX, for example, except for issues mandated by the language standards themselves, like arrangement of array data in memory. > UNIX has no calling standard. Of course it doesn't. It runs on heterogeneous hardware. Hardware drives calling standards. When particular Unix *implementations* don't have a single calling standard, that's the fault of the compiler writers, not the OS developers. > UNIX has no consistancy of library routines between languages. If I want > to do everything in C, I'm golden, but not everyone enjoys programming in > assembly language all the time. If I try to call system routines or RTL > routines from f77, I've got to pray there's some kind of wrapper > available, or WRITE IT MYSELF in C, and do my contortions to make it all > link properly. An implementation issue, which has nothing to do with Unix itself. > 3) lack of portability > > (3) So, if I take a binary from UNIX box X to UNIX box Y, it'll just > -work-, right? Y'know, like I can take a VMS binary from a 20-year-old > VAX and run it on the newest VAX. Ditto with VMS on an Alpha. OK. Now take that VMS binary and run it on a PC running one of the free VMS clones. Run it on the compatible platform sold by IBM, or Sun, or HP, or Bull, or Sequent, or DG, or any of N other vendors, which you've chosen because they're your preferred supplier or they gave you a package deal with an unrelated purchase or you like their support organization or they provided the best price/performance for your application mix or ... Conversely, take an AIX binary from a 320 running AIX 3.2.5 and try it on a rack system runinng 4.3. Take a binary from a Sun SPARCstation 10 and run it on a SPARCenter 2000. You stick with one platform and you get binary compatibility. You move between platforms and you recompile. VMS has nothing magic in this regard. (Well, actually it does, sort of - a microcoded CISC instruction set in the VAX implementation that in effect provides an extra layer of abstraction. There's nothing in Unix that prevents that, though. I don't know of a Unix platform with that sort of architecture, but that's unrelated to the design of Unix; it's an implementation decision.) Unix "portability" was oversold by a lot of people who didn't know what they were talking about, back in the 80's when the trade rags started getting interested in it. No question about that. Porting software from one Unix flavor to another can be a real pain. But that's an entirely different task than moving a program from one VMS machine to another, because VMS is not a multiplatform OS. The range of VMS machines is great, and a real VMS advantage, but they're all a single platform. (The Alpha VMS systems represented a major architecture change, but they were still completely under DEC control and designed for backward compatibility. They're no different in that respect than, say, CISC- and PowerPC-based AS/400s, which no one could reasonably claim are different platforms. Those are also binary-compatible, from the user's point of view.) > Okay, if you disallow that because they both come from the same vendor, > how about language compilers at least? If I build arbitrary program A on > an HP/UX box, will I be able to build it WITHOUT ANY CHANGES on AIX or AUX > or Digital UNIX? Yes, if it adheres to the appropriate standards. If not, why would anyone expect it to? And how is this an argument for VMS? Because there isn't a VMSalike OS that's not VMS, so you can't make the attempt in the first place? I'm sorry to see VMS apparently on the way out, too. There's definitely a place for proprietary, carefully controlled OSes. Unix is not suitable for every task; it's not even suitable for most of what computers are used for, I'd say. It's really best at what it's best known for, as a sandbox for programmers. That's an important function that it fulfills well. It's pretty lousy for people who just want to get some data crunching done and don't want to learn what a few hundred programmers each thought would be a good way to do things. But your comparisons between VMS and Unix mostly illustrate your inability or unwillingness to distinguish between hardware and software issues, or OS and applications, or core OS and third-party add-ons, or design and implementation, or hardware models and platforms. When people are posting religious war flamebait on alt.folklore.computers, they ought to at least strive for a bit of precision. Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University This is a "rubbering action game," a 2D platformer where you control a girl equipped with an elastic rope with a fishing hook at the end. -- review of _Umihara Kawase Shun_ for the Sony Playstation ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 02:49:14 GMT Organization: Micro Focus Inc. Lines: 43 Message-ID: <74kofa$8tf@news1.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: p-738.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news.tvd.be!newsfeed.wirehub.nl!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!207.172.3.37!feed1.news.rcn.net!rcn!pln-e!spln!extra.newsguy.com!newsp.newsguy.com!mww In article <74jjnt$i01$1@vixen.cso.uiuc.edu>, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) writes: > In article <366d1f6c.9761743@news.bright.net>, > wrote: > >"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. > >You need a shell program to process user input. > > Please point me to a system which is branded UNIX and does not contain > a Bourne shell. If you can't please take back your statement. The shell, Bourne or otherwise, isn't part of "straight Unix", if by that we mean the core OS. The shell has no special ties to the core OS, regardless of whether it was shipped with the OS or added later. The original poster's statement is perfectly valid. It's entirely possible to build a Unix system for specialized purposes that doesn't have a shell on it at all. (You'd need to replace those elements of second-tier, user-space processing that use the shell for convenience, like rc scripts, that you still wanted to use; but those aren't part of the core OS.) It would still be Unix. The core OS only requires a bootstrap loader, the kernel, probably some kernel helper processes like a page stealer (though some modern flavors build those into the kernel itself), init, and on modern platforms some shared library/object code. You would probably want some configuration programs like ifconfig (which can be run out of inittab) to do useful work; since you're not logging into the beast directly, you want it to be network-accessible. And then you want application code. None of those things should require a shell. Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University The lark is exclusively a Soviet bird. The lark does not like the other countries, and lets its harmonious song be heard only over the fields made fertile by the collective labor of the citizens of the happy land of the Soviets. -- D. Bleiman ###### From: skb@xmission.removethis.com (Scott Brown) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 03:48:59 GMT Organization: (none) Lines: 24 Message-ID: <366df034.14353548@news.xmission.com> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> <74kasb$47f$1@mathserv.mps.ohio-state.edu> NNTP-Posting-Host: slc722h.modem.xmission.com X-Trace: news.xmission.com 913175136 12876 166.70.7.214 (9 Dec 1998 03:45:36 GMT) X-Complaints-To: abuse@xmission.com NNTP-Posting-Date: 9 Dec 1998 03:45:36 GMT X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-was.dfn.de!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!xmission!nnrp.xmission!not-for-mail On 8 Dec 1998 22:57:15 GMT, dicks@math.ohio-state.NO.SPAM.edu (Ethan Dicks) wrote: >In article <366f6b01.9489096@Rockyd>, >Alexandre Pechtchanski wrote: >> >>And while we are at it, how in "hierarchical" Unix file system one can >>determine the (used) size of "/" partition (excluding other file systems >>hanging off) ? > >Under Solaris, du -d will avoid crossing filesystem boundaries. I typically >use "du -skd" to give me a report of the total size (not intermediate >directories), in K bytes, only on this filesystem. I also like "su -s *" >which tells me where all the space has gone underneath my current directory. I just use "df /". I'd personally love to have a VMS machine to play with. Aside from the fun of learning a new system, knowing a little about VMS would probably be worth something on a resume. Unfortunately, there doesn't seem to be a free VMS-alike that will run on a surplus 80486 PC. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 06:49:25 +0000 Organization: Tanglewood Message-ID: <366E1D75.F5D3585F@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74k850$996$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913193027 nnrp-03:29247 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 26 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Don Stokes wrote: > cp /mnt-pt/dir/file /mnt-pt2/newdir/newfile > > isn't really any different from: > > copy disk1:[dir]file disk2:[newdir]file Well, actually it is, because VMS doesn't allow wildcards in the device portion, whereas un*x will allow a recursive search to go from volume to volume. On this machine, which is quad-bootable (2 flavours of linux, OS/2 and M$dos), I find it useful to allow find to iterate through all the mounts when I really can't remember where something is. OTOH when I first moved from VMS to linux I kept wondering, "how do I *stop* things iterating through all the drives?". Basically, to my mind, it doesn't matter, both VMS and Linux allow me to perpitrate weird hacks which get the job done quickly, as does OS/2 to a lesser extent. The hacks just work in radically different ways. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 07:02:34 +0000 Organization: Tanglewood Message-ID: <366E208A.E2CCAB80@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913193028 nnrp-03:29247 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 24 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Howard S Shubs wrote: > Logical names are, in part, a way for system managers to alias things in > order to organize the system better. As a UNIX-head, you're familiar with > the concept of the "~" directory. Which is better, using that notation or > using the full path? The other virtue of logical names is that they can be different for different users. It is perfectly possible to set up the system in such a way that the system login file creates logical names for each session which depend on the user ID, time, phase of the moon, number of [mongooses|mongeese] in stock... The virtue of symlinks is that they are the same for all users, regardless of user ID... The two systems are complimentary, and implementing logical names in un*x, as well as symlinks, would be a Good Thing. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: deke.spamblock@generous.net Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 07:28:03 GMT Organization: GenerousCity is a virtue - find romance at http://generous.net Lines: 19 Message-ID: <366df4b1.44021400@news.bright.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: paul-cas1-cs-5.dial.bright.net X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newspeer1.nac.net!netnews.com!chnws02.mediaone.net!24.131.128.14!elnws02.ce.mediaone.net!24.131.1.12!denws01.mw.mediaone.net!news.altair.com!NewsNG.Chicago.Qual.Net!news.bright.net!not-for-mail On 8 Dec 1998 16:22:21 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >In article <366d1f6c.9761743@news.bright.net>, > wrote: >>"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. >>You need a shell program to process user input. > >Please point me to a system which is branded UNIX and does not contain >a Bourne shell. If you can't please take back your statement. > >-- Terry ------------------------ Let love find you! http://generous.net A list for flirting GenerousSingles-subscribe@onelist.com Over The Hill Gang GenerousSinglesOTHG-subscribe@onelist.com Personal Ads list GenerousProfiles-subscribe@onelist.com ###### From: deke.spamblock@generous.net Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 07:28:04 GMT Organization: GenerousCity is a virtue - find romance at http://generous.net Lines: 39 Message-ID: <366df4be.44034423@news.bright.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: paul-cas1-cs-5.dial.bright.net X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!howland.erols.net!netnews.com!chnws02.mediaone.net!24.131.128.14!elnws02.ce.mediaone.net!24.131.1.12!denws01.mw.mediaone.net!news.altair.com!NewsNG.Chicago.Qual.Net!news.bright.net!not-for-mail On 8 Dec 1998 16:22:21 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >In article <366d1f6c.9761743@news.bright.net>, > wrote: >>"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. >>You need a shell program to process user input. >Please point me to a system which is branded UNIX and does not contain >a Bourne shell. If you can't please take back your statement. OK, I will take back my statement. Obviously I am mistaken. One MUST use the Bourne shell. Those other shells are just figments of my imagination. Of course, why they call it the Bourne shell and not just "the command processor" like MS-DOS does, is beyond me. deke "It still moves." -- Galileo ###### From: Kevin Handy Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 10:42:42 -0700 Organization: Software Solutions, Inc Lines: 17 Message-ID: <366EB692.3CC46D52@srv.net> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> NNTP-Posting-Host: 12.7.216.58 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: 913224737 HLLAGTWDSD83AC 07C usenet52.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Mailer: Mozilla 4.04 [en] (Win95; I) To: David Given Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail David Given wrote: > Hey! How about an open source VMS clone based around the Linux kernel? > Look for the free-vms group. Hasn't gone very far yet, but if people start submitting code... http://www.cyclic.com/~kingdon/free-vms.html Seems to be heading toward VMS capability on Linux, although the stated desires are for a VMS on MACH clone. ###### From: Kevin Handy Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 10:43:26 -0700 Organization: Software Solutions, Inc Lines: 17 Message-ID: <366EB6BE.34B805C9@srv.net> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> NNTP-Posting-Host: 12.7.216.58 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: 913224782 HLLAGTWDSD83AC 07C usenet52.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Mailer: Mozilla 4.04 [en] (Win95; I) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail David Given wrote: > Hey! How about an open source VMS clone based around the Linux kernel? > Look for the free-vms group. Hasn't gone very far yet, but if people start submitting code... http://www.cyclic.com/~kingdon/free-vms.html Seems to be heading toward VMS capability on Linux, although the stated desires are for a VMS on MACH clone. ###### Date: 09 Dec 98 11:16:05 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <366df4be.44034423@news.bright.net> Message-ID: <1571.647T270T6763223@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 40 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 9 Dec 1998 12:26:05 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.208.190.2!news.globix.net!news-pen-2.sprintlink.net!news-east1.sprintlink.net!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.121 In article <366df4be.44034423@news.bright.net> deke.spamblock@generous.net (deke.spamblock) writes: >On 8 Dec 1998 16:22:21 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) >wrote: > > >In article <366d1f6c.9761743@news.bright.net>, > > wrote: > > > >>"Straight UNIX" doesn't do anything. It doesn't interface with the > >>user at all. You need a shell program to process user input. > > > >Please point me to a system which is branded UNIX and does not > >contain a Bourne shell. If you can't please take back your > >statement. > >OK, I will take back my statement. Obviously I am mistaken. One >MUST use the Bourne shell. Those other shells are just figments >of my imagination. No, one must be ABLE to use the Bourne shell. You can use any shell you want, but the Bourne shell must be one of your options. This assures that shell scripts written for the Bourne shell will run on any Unix box, even if you personally prefer ksh or whatever. >Of course, why they call it the Bourne shell and not just "the >command processor" like MS-DOS does, is beyond me. Because the Bourne shell is just of several shells (command processors, if you wish) that you can get for Unix. MS-DOS's COMMAND.COM, on the other hand, is "the command processor" simply because no replacement has really caught on. That's not to say you couldn't write one - but I'd want to be sure that COMMAND.COM is still available before I hand a .BAT file to a random user to run. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Date: 09 Dec 98 11:32:05 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: Message-ID: <821.647T270T6923295@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 21 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 9 Dec 1998 12:26:13 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.121 In article meowing@banet.net (Fluffy) writes: >> > Shell globbing is IMHO the most losing feature in Unix. >> >> That is your personal opinion. I Would say quite a number of Unix >> users doesn't agree with you here. As a former VMS user/programmer/ >> sysadm (between 1983 and 1987) I can say that VMS has its bright >> ideas, but having the globbing in the application is not one of them. > >Hmm. If Unix style shell globbing is a Good Idea, why do I so often >find myself having to rely on ugly hacks like xargs to run things >against large sets of files? Because it wasn't done quite right. It's still a Good Idea, though. Don't confuse a concept with its implementation. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 98 11:56:35 GMT Organization: UltraNet Communications, Inc. Lines: 40 Message-ID: <74lpss$9if$1@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <74j0l4$fc6$1@strato.ultra.net> NNTP-Posting-Host: d8.dial-15.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 9 Dec 1998 12:19:40 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d8 In article , alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >In article <74j0l4$fc6$1@strato.ultra.net> jmfbahciv@aol.com writes: > >>Rich? Have people changed the meaning of glob? Are they >>really talking about wildcarding as we in the -10 world >>knew and loved it? > >That's what it's used for, although it's not as efficient >as the Tops-20 method--I'm not sure how Tops-10 handled >this--since the glob() system call builds a list data >structure in memory pre-allocated by the program, where the GTJFN% >and GNJFN% JSYi chase things down through the directory data >block(s) (sort of like i-nodes for the Unix readers, but not >really). The TOPS-10 monitor just didn't do wildcarding. The CUSPs handled the wildcarding and did all that work. We had very strict rules about wildcarding syntax (and we could wildcard directories). It also generally only worked with a file system which is _not_ to be confused with a physical disk--something that people in this thread seem to be doing. > >It's the only meaning of "glob" I know--I only ever heard it >in a Unix context. What else does it refer to? It was a program shipped on the the TOPS-10 CUSP tape. Here's the intro to the blue revision of the phone book, 1971-- "GLOB reads multiple binary program files produced by MACRO and FORTRAN and generates an alphabetic cross-referenced list of all the global symbols encountered. It may also search specified file in Library Search Mode, checking for the globals only if the program was loaded by the LOADER in Library Search Mode." /BAH Subtract a hundred and four for e-mail. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 12:55:03 GMT Organization: Daedalus Consulting Lines: 59 Message-ID: <74lrv7$2p0$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74h82f$4mt@hermes.acs.unt.edu> <1998Dec8.101907.3786@lorelei.approve.se> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <1998Dec8.101907.3786@lorelei.approve.se>, Goran Larsson wrote: >If "VMS scores over" Unix by having the globbing in the application, thus It "scores" by having globbing in the *OS*, where anyone can get at it easily, shell, application, whatever. As opposed to unix where it's "roll your own", or reliant on a list passed as a limited size array in memory. >allowing the application to give different meaning to glob patterns, then >how is the filename completion code in the "shell" supposed to know how >to do a correct filename completion? The simple completion tasks like >completing a single filename is trivial, but what should be done if the >filename completion is attempted on a partial filename containing glob >patterns? The glob patterns are consistent to anything calling the RMS $SEARCH service (and that includes the LIB$FIND_FILE wrapper). And globbing works just fine thank you in DCLCOMPLETE, in that a filespec containing wildcards will be completed to the first file matching the wildcard. Which is more than I can say for any of the standard(ish) unix shells. I should also point out that this happens "for free" -- I didn't put any code in DCLCOMPLETE to handle globbing -- I did use $SEARCH (with a trailing '*' inserted in the file name to be completed), and $SEARCH just does the Right Thing. >Btw, how can you have spaces in VMS filenames today? You can't. I'm not sure if this is good or bad -- very few unix filenames have spaces in them in Real Life, and for good reason -- they're a nuisance. Same goes for any other "odd" characters that might show up that have other uses as delimiters or metacharacters -- just because you can use them doesn't mean you should. That said, there's no real reason the VMS character set is limited to the level it is, except the ODS1 ancestry that kept files in RADIX-50 format (three characters in one word), limiting the character set to 40 characters. ODS2 keeps its filenames in clear text. In fact, RMS is perfectly capable of handling file names contained in quotes, with spaces and other stuff (the quotes have to be included in the file name, but what goes between the quotes is pretty much up to the underlying ACP to deal with) -- I've quite frequently dealt with foreign tape formats (VMS is happy to deal with an ANSI labelled magtape as a "filesystem") and file names like MUA0:"File Name with Spaces". >> Shell globbing is IMHO the most losing feature in Unix. > >That is your personal opinion. I Would say quite a number of Unix >users doesn't agree with you here. As a former VMS user/programmer/ >sysadm (between 1983 and 1987) I can say that VMS has its bright >ideas, but having the globbing in the application is not one of them. Lets just say that shell globbing has certain limitations that programmers ignore at their peril. Application globbing (particularly if supported in a consistent fashion by the underlying OS) does not have these limitations. And let's keep the discussion to comparisons rather than "my OS is better than yours". -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### Message-ID: <366E74AB.6EC53A58@trailing-edge.com> From: Tim Shoppa Organization: Trailing Edge Technology X-Mailer: Mozilla 3.03Gold (X11; I; OpenVMS V7.0 DEC 3000 Model 300L) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 26 Date: Wed, 09 Dec 1998 13:01:31 -0400 NNTP-Posting-Host: 198.232.144.27 X-Trace: audrey2.cais.com 913226937 198.232.144.27 (Wed, 09 Dec 1998 13:08:57 EDT) NNTP-Posting-Date: Wed, 09 Dec 1998 13:08:57 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!in1.nntp.cais.net!199.0.216.204.MISMATCH!audrey2.cais.com!not-for-mail Eric J. Korpela wrote: > > In article , > Howard S Shubs wrote: > >In article <74jpov$4iu$1@agate.berkeley.edu>, > >korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: > > > >>The VMS file system has always felt to me like it is something cobbled onto > >>a flat file system. > > > >Strange. Probably due to a lack of experience with VMS vs some experience > >with RSX-11. > > Strange. Are you telling me that VMS directories weren't hacked onto a > flat file system? Only in the same sense that Unix directories were hacked onto a flat file system. With either, there's always a disk that's a "flat" collection of blocks, accessed by block number (whether you choose to call them LBN's, VBN's, or inodes.) (Yes, I know there are computers where mass disk storage is done associatively, and not by block number. But I don't know of any Unix or VMS systems that do so.) Tim. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 13:29:31 GMT Organization: Daedalus Consulting Lines: 86 Message-ID: <74ltvr$cfu$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <1998Dec8.105505.4519@lorelei.approve.se> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <1998Dec8.105505.4519@lorelei.approve.se>, Goran Larsson wrote: >How did you tell the completion routine how the application wanted the >filename completion to be done? If you have the files TEST.TXT and >TEST.EXE, then how did you handle this? > > TYPE TE > RUN TE Simple. It expands to "TEST". If you want the extension too, you type a dot and hit again. "RUN TEST" does what you want. VMS's parsing rules include the concepts of default filespecs and related filespecs. The default filespec usually just specifies a default file type, although it can have other components. So if you type "RUN TEST", the file will be parsed as "TEST.EXE" (getting the ".EXE" from the default). "TYPE TE" will also get you only "TYPE TEST", leaving you to type the ".TXT" (or ".T") yourself. Note however that if it was TEST.LIS that you wanted, well, ".LIS" is the default filespec for TYPE. Related filespecs are used to carry defaults over from command arguments, so an output file would have a default filespec of an appropriate file extension, and a related filespec of the input file. The default spec overrides fields specified in the related spec, which are in turn overridden by fields specified in the passed filename. > TYPE T*;-1 DCLCOMPLETE doesn't attempt to handle version numbers, since they're (a) short, and (b) not really typed in very often. (Damned handy when they are needed though!) But as I noted in a previous post, embedded wildcards are handled just fine by completing to a matching non-wildcard spec. >The Unix way is consistent, even if it makes no sense for some programs. >You know it is there and if you know how it works there will be no >surprices. The DCL way leaves the users at the mercy of the programmer, >you never know if the application handles it the way you expect. In the >end it is just a user preference. If all users wanted it the DCL way, >then by now we would have a number of Unix shells doing it that way, or? Since VMS provides one set of globbing rules, the result is *extremely* consitent. Like really really consistent. There's really only one way to do it. I've never had a surprise from wildcard expansion. Folks might not *like* the way $SEARCH parses filenames, but at least they get to *consistently* dislike it... >$ command **/*.c All files matching *.c >$ command **/*.c(u:hoh:) ...and owned by user 'hoh' >$ command **/*.c(m-2) ...and modified within the last 2 days And zsh is the standard shell in exactly which breed of unix? You were saying something about consistency? 8-) (This reminds me a bit about RSTS/E filenames -- you could put switches on the filename, and file protections to affect the way the file was opened and handled. Kinda neat, if you didn't mind 6.3 letter filenames in [p,pn] directories... 8-) >Note that in all these examples I can press TAB instead of RETURN and >have the pattern replaced by whatever it expands to, e.g. > >$ ls **/c*(m-2) >$ ls config.cache config.h config.log config.status src/crc.h > >The shell (zsh) replaced the command line with the expanded command line >for me. I have no idea how this could be implemented in DCL. DCLCOMPLETE is pretty simpleminded (I wrote it for no other reason than because some idiot said it couldn't be done), but a hypothetical future version *could* look at the DCL command table, determine if the particular command element permits a list of files and complete the files as "file,file,file". But there's no real reason to do so -- any application that takes a list will parse the list properly anyway. >Make that > >$ command `find . -name='*.c' -print` > >to be portable as POSIX requires that -print be given if intended. Yeah, yeah. I use BSD which tends to assume that real people use the damn thing, not just standards weenies. 8-) -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 13:42:30 GMT Organization: Daedalus Consulting Lines: 31 Message-ID: <74luo6$frj$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74k6sf$dr5$1@agate.berkeley.edu>, Eric J. Korpela wrote: >In article <74jv4i$t2a$1@vixen.cso.uiuc.edu>, >TERENCE MURPHY wrote: >> >>>Well, it wouldn't be that difficult to create the equivalent of logical >>>names in a unix shell. >> >>It would be easy to do in the shell but it would require a complete >>re-design and re-write of every program. > >I think you miss the point entirely. Like wildcard expansion, logical >names would be expanded by the shell, not by the programs. In fact every >unix system I've used has shells that support the logical name "~" which >in VMS speak would be "HOME:". The programs don't need to know how to >expand "~" because they don't see it. The standard Bourne shell doesn't understand '~'. On the other hand, SYS$LOGIN: is the user's login directory. So the application can reference SYS$LOGIN:file.typ any time and know it's going to get file.typ the login directory. It doesn't have to translate SYS$LOGIN: to find the directory -- it's done by the OS. And it's done for file names hard coded in the application, passed from DCL, typed in by the user or obtained from a configuration file or option somewhere. '~' is only parsed by the shell -- if you want to use '~' in a file name passed some other way, the application has to be smart enough to do the lookup itself. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 9 Dec 1998 14:08:28 +0100 Organization: [Posted via] Interactive Networx Lines: 15 Message-ID: <1djrqt0.1k1zbkajihk00N@[10.0.0.3]> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> <74kc13$chj$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: n37-24.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.nacamar.de!news-hh.maz.net!unlisys!news.snafu.de!jnickelsen TERENCE MURPHY wrote: > logicals need to be expanded in interactive programs, not just the shell. Logicals are, as someone else pointed out, expanded by the file system layer of the operationg system (RMS). This way *any* program uses logicals -- whether the author intended that or not. I once had to use a program that read input parameters only from a file with a predefined filename. PITA. Well, wrap some DCL script around it and make this filename a logical -- voila, the filename to read from can be made a parameter of the script. -- Juergen Nickelsen ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 9 Dec 1998 14:08:31 +0100 Organization: [Posted via] Interactive Networx Lines: 18 Message-ID: <1djrsod.6w627whz8fvgN@[10.0.0.3]> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: n37-24.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!newsfeed.ecrc.net!newsfeed.nacamar.de!news-hh.maz.net!unlisys!news.snafu.de!jnickelsen Howard S Shubs wrote: [Unix] > "Well, this directory contains files x y and z unless -that- partition is > mounted, in which case it contains files a b and c. So, how do you get at > x y and z to append them to a b and c? Gotta use a scratch area. Yeah, > that's smooth." Mounting a file system over a non-empty directory just isn't done usually, because it is, as you point out, usually a stupid idea. I can see three reasons why it might be done: - Someone is stupid in general. - Someone is not stupid in general, but making a stupid mistake. - Someone wants precisely that: shadow a set of one files with a set of different files. This might be useful in some situations. -- Juergen Nickelsen ###### From: meowing@banet.net (Fluffy) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> <1998Dec8.101907.3786@lorelei.approve.se> Organization: If one may meow, all may meow. X-P-Meow: Meow Mail-Copies-To: never X-URL: http://members.tripod.com/~gerglery/ Cancel-Lock: sha1:+JJ5gqNMrWa1NX2EejPUh6e+fT8= Message-ID: X-Newsreader: slrn (0.9.5.3-canlock UNIX) Cancel-Lock: sha1:+JJ5gqNMrWa1NX2EejPUh6e+fT8= NNTP-Posting-Host: 32.100.243.104 Date: 9 Dec 1998 15:53:08 GMT X-Trace: 9 Dec 1998 15:53:08 GMT, 32.100.243.104 Lines: 13 X-Notice: Items posted that violate the IBM.NET Acceptable Use Policy X-Notice: should be reported to postmaster@ibm.net X-Complaints-To: postmaster@ibm.net Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsfeed.cwix.com!165.87.194.242!newsm2.ibm.net!ibm.net!news3.ibm.net!32.100.243.104 > > Shell globbing is IMHO the most losing feature in Unix. > > That is your personal opinion. I Would say quite a number of Unix > users doesn't agree with you here. As a former VMS user/programmer/ > sysadm (between 1983 and 1987) I can say that VMS has its bright > ideas, but having the globbing in the application is not one of them. Hmm. If Unix style shell globbing is a Good Idea, why do I so often find myself having to rely on ugly hacks like xargs to run things against large sets of files? -- "FEAST!" --Alice ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 16:20:47 GMT Organization: University of Illinois at Urbana-Champaign Lines: 20 Message-ID: <74m80v$4od$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366f6b01.9489096@Rockyd> <74kasb$47f$1@mathserv.mps.ohio-state.edu> <366df034.14353548@news.xmission.com> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <366df034.14353548@news.xmission.com>, Scott Brown wrote: >I'd personally love to have a VMS machine to play with. Aside from >the fun of learning a new system, knowing a little about VMS would >probably be worth something on a resume. > >Unfortunately, there doesn't seem to be a free VMS-alike that will run >on a surplus 80486 PC. Yes, but there is free VMS for VAX (peecee's are boring as hell anyways, why would you want to use one?) Just go to www.ebay.com, pick yourself up a VAXstation (for $200), then go to www.montagar.com and pick up a VMS license & media (for $30), and you'll be all set for the best computing experience of your life. Much more fun that Linux at least. -- Terry ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 16:34:55 GMT Organization: University of Illinois at Urbana-Champaign Lines: 53 Message-ID: <74m8rf$5ev$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <74kofa$8tf@news1.newsguy.com>, Michael Wojcik wrote: >The shell, Bourne or otherwise, isn't part of "straight Unix", if >by that we mean the core OS. The shell is part of POSIX.2 (the second and third letters stand for "operating system") and also part of the OpenGroup Unix specification. If you delete /bin/sh what you have left will not be Unix (since it will not pass the POSIX or OpenGroup specifications). >The shell has no special ties to the core OS, regardless of whether it >was shipped with the OS or added later. The original poster's >statement is perfectly valid. It's -part- of the OS. Unless you don't understand the difference the kernel and the OS. >It's entirely possible to build a Unix system for specialized >purposes that doesn't have a shell on it at all. (You'd need >to replace those elements of second-tier, user-space processing >that use the shell for convenience, like rc scripts, that you >still wanted to use; but those aren't part of the core OS.) It >would still be Unix. I am signficantly aware of this. At work I'm running a bare executable on top of the Linux kernel (as an embedded system)-- no shell. This has absolutely NOTHING to do with Unix though -- it certainly would not be branded Unix by the OpenGroup or be POSIX compliant (and not because it's Linux, but because it doesn't have a shell and other core utilities). >The core OS only requires a bootstrap loader, the kernel, probably >some kernel helper processes like a page stealer (though some modern >flavors build those into the kernel itself), init, and on modern >platforms some shared library/object code. You would probably want >some configuration programs like ifconfig (which can be run out of >inittab) to do useful work; since you're not logging into the beast >directly, you want it to be network-accessible. And then you want >application code. > >None of those things should require a shell. A massive quantity of off-the-shelf application which effectively say "requires (e.g.) Solaris to run" require not just the kernel, but lots of helper programs (of which the most common is probably the shell). Please note that they say "requires Solaris" not requires "Solaris plus /bin/sh and /bin/ls and /bin/cp", etc. Therefore one can conclude that the term "Solaris" means not the kernel but the kernel as well as included utilities. It's like this for any Unix. -- Terry ###### From: bjh21@cus.cam.ac.uk (Ben Harris) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 16:58:13 GMT Organization: University of Cambridge, England Lines: 14 Message-ID: <74ma75$eno$1@pegasus.csx.cam.ac.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74kasb$47f$1@mathserv.mps.ohio-state.edu> <366df034.14353548@news.xmission.com> <74m80v$4od$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: taurus.cus.cam.ac.uk Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!rill.news.pipex.net!pipex!server1.netnews.ja.net!pegasus.csx.cam.ac.uk!bjh21 In article <74m80v$4od$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: >Yes, but there is free VMS for VAX (peecee's are boring as hell >anyways, why would you want to use one?) Just go to www.ebay.com, pick >yourself up a VAXstation (for $200), then go to www.montagar.com and >pick up a VMS license & media (for $30), ... assuming you're a member of one of the relevant DECUS chapters. As far as I can see, that'd entail my moving to the Netherlands at least, which is probably more effort than I'd like to go to for a VMS licence. -- Ben Harris Computer Officer, Corpus Christi College, Cambridge. ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 17:00:44 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 19 Message-ID: <74mabs$rrl$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article , Howard S Shubs wrote: >In article <74jpov$4iu$1@agate.berkeley.edu>, >korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: > >>The VMS file system has always felt to me like it is something cobbled onto >>a flat file system. > >Strange. Probably due to a lack of experience with VMS vs some experience >with RSX-11. Strange. Are you telling me that VMS directories weren't hacked onto a flat file system? Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: Ariel Scolnicov Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 09 Dec 1998 17:04:02 +0200 Organization: NetVision Israel Lines: 39 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> <74luo6$frj$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: bioserv.compugen.co.il X-Trace: news.netvision.net.il 913215922 11916 194.90.227.153 (9 Dec 1998 15:05:22 GMT) X-Complaints-To: abuse@netvision.net.il NNTP-Posting-Date: 9 Dec 1998 15:05:22 GMT X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-a.ais.net!ais.net!news.idt.net!news-peer.gip.net!news-dc.gip.net!news.gsl.net!gip.net!news-feed.netvision.net.il!194.90.1.15.MISMATCH!news!not-for-mail don@news.daedalus.co.nz (Don Stokes) writes: [...] > On the other hand, SYS$LOGIN: is the user's login directory. So the > application can reference SYS$LOGIN:file.typ any time and know it's going > to get file.typ the login directory. It doesn't have to translate > SYS$LOGIN: to find the directory -- it's done by the OS. And it's > done for file names hard coded in the application, passed from DCL, > typed in by the user or obtained from a configuration file or option > somewhere. '~' is only parsed by the shell -- if you want to use '~' > in a file name passed some other way, the application has to be smart > enough to do the lookup itself. On the other hand, $HOME (the environment variable named HOME) is the user's login directory. So the application can reference "$ENV{HOME}/file.typ" (if it's in Perl; substitute your favourite string-processing language's idiom) any time and know it's going to get file.typ [in] the login directory. So it's *not* an issue when hard-coding filenames in an application. The problems are when you want the user to enter a filename interactively and have the application translate ~someOtherUser/file.typ to file.typ in someOtherUser's home directory. This is really a csh thing (which indeed expands ~, unlike say sh), not a Unix thing. Unix bigots may (rightfully?) say "So what? My tcsh also expands =n to the n'th directory on its directory stack -- you want programs to do that too? What about expanding !$ to the last argument of the previous command?". It's a shell abbreviation, not a system thing (can you run {z,tc,ba,...}sh on anything unUnixy?). This is good because your favourite shell can add your favourite abbreviations, and it's bad because the very best abbreviations (like ~user) get to be treated (by users) as alternate filenames, while applications don't necessarily know of it. -- Ariel Scolnicov /---------------\ "GCAAGAATTGAACTGTAG" Compugen Ltd. |Join the Dark | Tel: +972-2-6795059 (Jerusalem) 17 Hamacabim St. |Side! Just Say:| Tel: +972-3-9348482 (Main office) Petakh-Tikva, 49220 | "Yes, but..."| Fax: +972-3-9348489 ISRAEL \---------------/ ariels@compugen.co.il ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 17:22:49 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 34 Message-ID: <74mbl9$8j$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74hkig$oaq$1@news.wlg.netlink.net.nz> <1998Dec8.105505.4519@lorelei.approve.se> <74ltvr$cfu$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74ltvr$cfu$1@news.wlg.netlink.net.nz>, Don Stokes wrote: > >>$ command **/*.c All files matching *.c >>$ command **/*.c(u:hoh:) ...and owned by user 'hoh' >>$ command **/*.c(m-2) ...and modified within the last 2 days > >And zsh is the standard shell in exactly which breed of unix? You were >saying something about consistency? 8-) There's no such thing as "the standard shell". All unix systems will have the bourne shell or a compatible shell, to allow compatibility of shell scripts. 95% of systems will also have the C shell or a compatible shell. Any other shells are up to the administrator or a user to install. (Yes, a user can install their own shell. Not as a login shell, but it's easy to run the shell of your choice from your login files). Every shell script starts with a line telling the system which shell to run and which shell options to use, so you don't end up with script confusion. And the point of having a standard shell is? So everyone can be hobbled by being restricted to DCL for their command line? Don't you just love how easy it is to pipe the output of one program to the input of another under VMS? Don't you just love how VMS programs all have the /PAGE option so their output doesn't scoll off the screen. You were saying something about consistency? Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: ejb@klamath.demon.co.uk (Edward John M. Brocklesby) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 18:10:54 GMT Message-ID: <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: klamath.demon.co.uk X-NNTP-Posting-Host: klamath.demon.co.uk:212.228.198.242 X-Trace: news.demon.co.uk 913227054 nnrp-08:18505 NO-IDENT klamath.demon.co.uk:212.228.198.242 X-Complaints-To: abuse@demon.net X-Newsreader: slrn (0.9.4.3 UNIX) Lines: 22 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!klamath.demon.co.uk!ejb On about Mon, 07 Dec 1998 10:07:15 -0600, Howard S Shubs might have written: >UNIX... no wonder it's being supplanted by Windows NT. And the fact that >VMS appears to be ported -from- more often than it's ported -to-, is >nothing less than a major tragedy. We're porting our IRC bot (Ilona) to VMS at some point. (It might well be me who'll have to do this). But I can't find any documentation for it.. The system I'll be using has MULTINET TCP/IP installed, but I can't find anything like section 3 on Unix .. I suppose I just haven't looked hard enough :) Are there any manuals available online? >-- >Howard S Shubs The Denim Adept -- Edward John M. Brocklesby System Administrator, Klamath Public Access Unix System | Data: +44 1865 454802 | Free-VMS - Expand your mind - www.free-vms.org | Free UNIX Accounts | Finger ejb@deep-thought.ml.org for PGP key ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 18:27:53 GMT Organization: Laser-Scan Ltd. Lines: 19 Message-ID: <74mff9$msa@relay.lsl.co.uk> References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!news-lond.gip.net!news.gsl.net!gip.net!diablo.theplanet.net!join.news.pipex.net!pipex!news.lsl.co.uk!benh lisard@zetnet.co.uk wrote: : Someone please tell me if I'm wrong (and a quick description of how : would also be nice ;> ) but I've been brought up to believe that the AT : and above use the INSB instruction to do 512 reads from a port per : sector, and getting them to do it with DMA is a bit of a nightmare. You are wrong. IDE drives and PC IDE interfaces have supported DMA for several years now. I believe the bus protocol used for this may be a little different from the original one (essentially a subset of ISA). The timing is certainly different. : And certainly the PC BIOS blocks on read. Woof. Yes, the PC BIOS only supports synchronous I/O. That is why no real OS for the PC uses the BIOS for I/O once it has booted. -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: jnickelsen@acm.org's message of Wed, 9 Dec 1998 14:08:31 +0100 Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> Date: Wed, 9 Dec 1998 19:37:17 GMT Lines: 27 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!newsfeed.enteract.com!ix.netcom.com!netcom!alderson In article <1djrsod.6w627whz8fvgN@[10.0.0.3]> jnickelsen@acm.org (Juergen Nickelsen) writes: >Mounting a file system over a non-empty directory just isn't done usually, >because it is, as you point out, usually a stupid idea. I can see three >reasons why it might be done: > - Someone is stupid in general. > - Someone is not stupid in general, but making a stupid mistake. > - Someone wants precisely that: shadow a set of one files with a set > of different files. This might be useful in some situations. That third reason is important. An example: On HP-UX systems, there is a system management tool called SAM, the binaries for which live in a directory on /usr. Now, on any decent system, /usr is going to live on a different filesystem than /, so in single-user mode it may not be available--like when moving /usr to a larger filesystem. So I, at least, put a copy of /usr/sam (and a couple of other tools) in the /usr directory which is the mountpoint for the /usr filesystem. Made my life a lot simpler, and I hardly think it was stupid--it became SOP for a large (then $750,000,000) networking company I worked for at the time... -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 20:31:29 GMT Lines: 50 Message-ID: <74mmn1$4l7$3@irk.zetnet.co.uk> References: <74jjnt$i01$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: man-178.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 913235489 4775 194.247.40.226 (9 Dec 1998 20:31:29 GMT) NNTP-Posting-Date: 9 Dec 1998 20:31:29 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!diablo.theplanet.net!news.freedom2surf.net!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-08 tsmurphy@cs.uiuc.edu(TERENCEMURPHY) said: : wrote: :>"Straight UNIX" doesn't do anything. It doesn't interface with the :>user at all. You need a shell program to process user input. :Please point me to a system which is branded UNIX and does not :contain a Bourne shell. If you can't please take back your :statement. Doesn't the lack of a Unix system without a shell rather reinforce his statement? Doesn't it emphasise the point that Unix is not synonymous with what your /etc/passwd entry says your shell is? The point is that shells are optional (although you might have a job removing one that isn't running); all the shells do is expose the kernel's abilities in a pleasant (and interactive) way. Some people use GNU Emacs as their shell. *shudder* Does that mean GNU Emacs is a part of Unix, in the inseparable sense you seem to imply above? The whole point of Unix is that there *is* no "base shell"; they're just other programs. The Bourne shell was just the first, and remains the least capable and most ubiquitous. It occupies a similar position to vi in that respect. But there are a huge number of others. Moreover, they don't even do anything special; 'ps' is more privileged than 'sh' (it has to read process information, so it's a suid). deke commented that writing a shell is "not exactly rocket science". He's right. You can knock one up in a few lines of C, if you so desire; the only functions you really need are fgets, fprintf, fork, and exec. No problem. And guess what? it subsequently has exactly the same status in the system as sh, csh, bash, et al. (The Bourne shell does have some advantages; it's a fairly lightweight scripting platform, and it keeps out of the way as much as possible. But I still don't think it's actually meant to be *used* by anyone. ;> ) If you're going to criticise Unix, it's as well to do so from a position of understanding. I still haven't seen anything from you that suggests you have any deep understanding of the system at all, and I'm a neophyte who'll openly admit a limited knowledge. The whole point of Unix is that if you don't like it, you can change it. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing In fact, I might just have thought of one; the old ICL thing I used to have in my secondary school. I was told it ran on Unix beneath everything, but the "everything" was an office package from hell. There might have been a shell somewhere, but I'm not sure anyone ever used it (which surely counts). It was about 10 years ago now, but it was surely UNIX-branded then. ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 20:31:33 GMT Lines: 27 Message-ID: <74mmn5$4l7$4@irk.zetnet.co.uk> References: NNTP-Posting-Host: man-178.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 913235493 4775 194.247.40.226 (9 Dec 1998 20:31:33 GMT) NNTP-Posting-Date: 9 Dec 1998 20:31:33 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-07 hshubs@mindspring.com(HowardSShubs) said: :Case-sensitivity as a GOOD thing? Short, arbitrary combinations of :letters for normal commands?? Horseshit. All that should be :redone to make it easier to use. Keep the older commands for :compatibility's sake, sure, but start working on fixing the :problems. Well, you can rename 'ls' to 'dir' if you prefer. ;> But I'll agree that case sensitive filenames are seriously broken; I'm forever getting stung by that one. Sure, keep the case of creation, but don't insist on it! (On the other hand, C was case-sensitive where BCPL was not, so perhaps they're just used to it in New Jersey. ;> ) :>It runs fins. Agreed, both systems are BSD based and they are :>both for the same architecture. No, it wont run on my Sun. :>Curious, will your VAX binary run on the alpha (I'd guess yes)? :>How about the other direction? :Neither, but a recompile/relink sequence will almost always succeed. I'm not surprised. There aren't generally too many variables in a single vendor system. Unix has never been a single vendor system, not even before its commercial life (eg. BSD). -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 20:31:37 GMT Lines: 20 Message-ID: <74mmn9$4l7$5@irk.zetnet.co.uk> References: NNTP-Posting-Host: man-178.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 913235497 4775 194.247.40.226 (9 Dec 1998 20:31:37 GMT) NNTP-Posting-Date: 9 Dec 1998 20:31:37 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-07 alderson@netcom.com said: :Is globbing done *in the application*, or is it done by the :operating system via a (set of) system call(s) as in Tops-20? I :realize the 32-bit weenies had a serious case of NIH, but did they :really get that wrong, too? It sounds, from what people have written here, as if globbing is done on VMS by a set of system calls. MS-DOS does it that way too, although DOS' globbing is seriously broken (a*b.c will find all the files that match a*.c, since * gets expanded to ???????? - Win32 fixes it for long names). Unix, on the other hand, *does* do globbing in the application. The shell is the application that does the globbing; the idea is that nothing else then has to worry about it. On the other hand, since a directory is merely a list of filenames and i-nodes, you can open a directory as a file, which eases the pain somewhat. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: "Andrew Cadley" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 9 Dec 1998 20:34:25 -0000 Organization: University of East Anglia, Norwich, Norfolk, NR47TJ, UK Lines: 11 Message-ID: <74mmvu$1lf@cpca3.uea.ac.uk> References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> <74mff9$msa@relay.lsl.co.uk> NNTP-Posting-Host: oak6d.vlg1.uea.ac.uk Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Newsreader: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!nntp.news.xara.net!xara.net!rill.news.pipex.net!pipex!server1.netnews.ja.net!server2.netnews.ja.net!cpca3.uea.ac.uk!news Ben Hutchings wrote in message <74mff9$msa@relay.lsl.co.uk>... :Yes, the PC BIOS only supports synchronous I/O. That is why no real :OS for the PC uses the BIOS for I/O once it has booted. So what about Windows then? ;-) AndyC ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 20:35:46 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 30 Message-ID: <366ebdd2.4345687@news.vip.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <366df4be.44034423@news.bright.net> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.18 X-Trace: 913235589 A01OARAUVD412CCD1C usenet58.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail deke.spamblock@generous.net wrote: >On 8 Dec 1998 16:22:21 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: > > >In article <366d1f6c.9761743@news.bright.net>, > > wrote: > >>"Straight UNIX" doesn't do anything. It doesn't interface with the user at >all. > >>You need a shell program to process user input. > > >Please point me to a system which is branded UNIX and does not contain > >a Bourne shell. If you can't please take back your statement. > >OK, I will take back my statement. Obviously I am mistaken. One MUST use the >Bourne shell. Those other shells are just figments of my imagination. > >Of course, why they call it the Bourne shell and not just "the command >processor" like MS-DOS does, is beyond me. "Bourne shell": two syllables. "command processor": rather more. What's this "just"? Oh, you want a boring name. Fine, "shell". Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: meowing@banet.net (Fluffy) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 20:49:33 GMT Organization: If one may meow, all may meow. Message-ID: References: <821.647T270T6923295@sky.bus.com> X-Trace: mail2news.demon.co.uk 913236695 mail2news:4078 mail2news mail2news.demon.co.uk X-Complaints-To: abuse@demon.net X-Mail2News-Path: news.demon.net!out2.ibm.net!meowhost.meow.invalid X-P-Meow: Meow Mail-Copies-To: never X-URL: http://members.tripod.com/~gerglery/ Cancel-Lock: sha1:pBAZmKTCW8WoFPbDPR2E6HY/+hk= X-Newsreader: slrn (0.9.5.3-canlock UNIX) Lines: 15 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!diablo.theplanet.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!mail2news.demon.co.uk!not-for-mail Charlie Gibbs wrote: > meowing@banet.net (Fluffy) writes: > >Hmm. If Unix style shell globbing is a Good Idea, why do I so often > >find myself having to rely on ugly hacks like xargs to run things > >against large sets of files? > > Because it wasn't done quite right. It's still a Good Idea, though. But when that filespac is just used a a param called by a function at runtime, the problem is solved. And that's what VMS lets you do. > Don't confuse a concept with its implementation. I haven't. ###### From: Sam Merritt Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> User-Agent: tin/pre-1.4-981114 ("The Watchman") (UNIX) (FreeBSD/2.2.6-RELEASE (i386)) Lines: 20 Message-ID: <0WBb2.85709$gI1.17874@news2.giganews.com> Date: Wed, 09 Dec 1998 21:26:20 GMT NNTP-Posting-Host: 209.198.128.35 X-Trace: news2.giganews.com 913238780 209.198.128.35 (Wed, 09 Dec 1998 15:26:20 CDT) NNTP-Posting-Date: Wed, 09 Dec 1998 15:26:20 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!207.207.0.26!nntp.giganews.com!news2.giganews.com.POSTED!not-for-mail TERENCE MURPHY wrote: > In article <366d1f6c.9761743@news.bright.net>, > wrote: >>"Straight UNIX" doesn't do anything. It doesn't interface with the user at all. >>You need a shell program to process user input. > Please point me to a system which is branded UNIX and does not contain > a Bourne shell. If you can't please take back your statement. *yawn* OS/390 Unix http://www.s390.ibm.com/oe/portbk/bpxacenv.html Other shells, such as C, Bourne, and bash, are not available. -- Sam Hayes Merritt, III http://www.frenzy.com/~harter/ ###### From: ejb@klamath.demon.co.uk (Edward John M. Brocklesby) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 21:44:03 GMT Message-ID: <913239843.29831.1.nnrp-05.d4e4c6f2@news.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> <74luo6$frj$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: klamath.demon.co.uk X-NNTP-Posting-Host: klamath.demon.co.uk:212.228.198.242 X-Trace: news.demon.co.uk 913239843 nnrp-05:29831 NO-IDENT klamath.demon.co.uk:212.228.198.242 X-Complaints-To: abuse@demon.net X-Newsreader: slrn (0.9.4.3 UNIX) Lines: 40 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!woodstock.news.demon.net!demon!news.demon.co.uk!demon!klamath.demon.co.uk!ejb On about 09 Dec 1998 17:04:02 +0200, Ariel Scolnicov might have written: >don@news.daedalus.co.nz (Don Stokes) writes: >[...] >> On the other hand, SYS$LOGIN: is the user's login directory. So the >> application can reference SYS$LOGIN:file.typ any time and know it's going >> to get file.typ the login directory. It doesn't have to translate >> SYS$LOGIN: to find the directory -- it's done by the OS. And it's >> done for file names hard coded in the application, passed from DCL, >> typed in by the user or obtained from a configuration file or option >> somewhere. '~' is only parsed by the shell -- if you want to use '~' >> in a file name passed some other way, the application has to be smart >> enough to do the lookup itself. > >On the other hand, $HOME (the environment variable named HOME) is the >user's login directory. So the application can reference >"$ENV{HOME}/file.typ" (if it's in Perl; substitute your favourite >string-processing language's idiom) any time and know it's going to >get file.typ [in] the login directory. You can't do this in C, and most popular Unix language. >system thing (can you run {z,tc,ba,...}sh on anything unUnixy?). This Yes. AmigaOS. And ~ even works there. >-- >Ariel Scolnicov /---------------\ "GCAAGAATTGAACTGTAG" >Compugen Ltd. |Join the Dark | Tel: +972-2-6795059 (Jerusalem) >17 Hamacabim St. |Side! Just Say:| Tel: +972-3-9348482 (Main office) >Petakh-Tikva, 49220 | "Yes, but..."| Fax: +972-3-9348489 >ISRAEL \---------------/ ariels@compugen.co.il -- Edward John M. Brocklesby System Administrator, Klamath Public Access Unix System | Data: +44 1865 454802 | Free-VMS - Expand your mind - www.free-vms.org | Free UNIX Accounts | Finger ejb@deep-thought.ml.org for PGP key ###### From: Mike Swaim Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> <74mff9$msa@relay.lsl.co.uk> <74mmvu$1lf@cpca3.uea.ac.uk> Organization: PointeCom User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/2.2.8-RELEASE (i386)) Lines: 16 Message-ID: Date: Wed, 09 Dec 1998 21:50:27 GMT NNTP-Posting-Host: 209.127.0.130 X-Trace: news2.giganews.com 913240227 209.127.0.130 (Wed, 09 Dec 1998 15:50:27 CDT) NNTP-Posting-Date: Wed, 09 Dec 1998 15:50:27 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.wirehub.nl!newspeer1.nac.net!newspeer.monmouth.com!nntp.giganews.com!news2.giganews.com.POSTED!gemini.c-com.net!swaim Andrew Cadley wrote: : Ben Hutchings wrote in message <74mff9$msa@relay.lsl.co.uk>... : :Yes, the PC BIOS only supports synchronous I/O. That is why no real : :OS for the PC uses the BIOS for I/O once it has booted. : So what about Windows then? ;-) AFAIK, it only uses the BIOS if something about the disks confuses it, in which case it drops into "compatibility mode." I believe that Windows for Workgroups was the first version of Windows to bypass the BIOS for disk I/O. -- Mike Swaim, Avatar of Chaos: Disclaimer:I sometimes lie. Home: swaim@c-com.net Alum: swaim@alumni.rice.edu Quote: "Boingie"^4 Y,W&D ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 21:53:11 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 18 Message-ID: <74mrg7$ach$1@agate.berkeley.edu> References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> <74mff9$msa@relay.lsl.co.uk> <74mmvu$1lf@cpca3.uea.ac.uk> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74mmvu$1lf@cpca3.uea.ac.uk>, Andrew Cadley wrote: >Ben Hutchings wrote in message <74mff9$msa@relay.lsl.co.uk>... > >:Yes, the PC BIOS only supports synchronous I/O. That is why no real >:OS for the PC uses the BIOS for I/O once it has booted. > >So what about Windows then? ;-) I won't claim Windows is a real OS, but not even it uses the BIOS for disk I/O. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:12:39 -0600 Organization: MindSpring Enterprises Lines: 22 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <366d8031.0@news2.uswest.net> <367591ad.19390965@Rockyd> <366dabe5.0@news3.uswest.net> <74n0tb$i8i$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: d1.56.d0.6a X-Server-Date: 10 Dec 1998 04:09:21 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!204.59.152.222!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74n0tb$i8i$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz (Don Stokes) wrote: >DIRECTORY can tell you exactly how much disk space a file takes -- the >/SIZE qualifier can be /SIZE=USED, /SIZE=ALLOCATION or /SIZE=ALL. >ALLOCATION gives you the actual blocks used, while USED gives the blocks >actually holding data. VMS makes no attempt to hide the "overhead", in >fact every block of the file system is mapped to a file, including the >home block and backup home blocks, file headers, storage bitmap, boot >blocks, bad blocks and so-on. Thus, the grand total of >DIR/GRAND/SIZE=ALLOC [*...] plus the free space as reported by SHOW DEVICE >should[1] exactly match the size of the disk. Unless things have changed, this is not quite true. DIR/SIZ=ALL tells you the number of disk blocks the file takes, yes. However, if you've got disk quotas enabled, you'll find that adding up all the allocated blocks will not jibe with the number given by SHOW QUOTA. This is because not only does the quota charge the file blocks against you, but also the file header blocks for those files you own. Each file as at least one file header block. Some have more, for various reasons. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:15:19 -0600 Organization: MindSpring Enterprises Lines: 11 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <1998Dec8.105505.4519@lorelei.approve.se> <74ltvr$cfu$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: d1.56.d0.6a X-Server-Date: 10 Dec 1998 04:12:00 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74ltvr$cfu$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz (Don Stokes) wrote: >Since VMS provides one set of globbing rules, the result is *extremely* >consitent. Like really really consistent. There's really only one >way to do it. I've never had a surprise from wildcard expansion. Agreed. If some program does it some other way, the programmer(s) involved should be forced to fix it. -- Howard S Shubs The Denim Adept ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 22:18:29 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 35 Message-ID: <74msvl$c2k$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74luo6$frj$1@news.wlg.netlink.net.nz> <913239843.29831.1.nnrp-05.d4e4c6f2@news.demon.co.uk> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <913239843.29831.1.nnrp-05.d4e4c6f2@news.demon.co.uk>, Edward John M. Brocklesby wrote: >> >>On the other hand, $HOME (the environment variable named HOME) is the >>user's login directory. So the application can reference >>"$ENV{HOME}/file.typ" (if it's in Perl; substitute your favourite >>string-processing language's idiom) any time and know it's going to >>get file.typ [in] the login directory. > >You can't do this in C, and most popular Unix language. Huh? #include #include #include FILE *open_file_in_HOME(char *file,char *type) { /* Warning! Poor error checking... */ char filename[_POSIX_PATH_MAX]; char *home=getenv("HOME"); if (home) { strncpy(filename,getenv("HOME"),_POSIX_PATH_MAX); strncat(filename,"file.typ",_POSIX_PATH_MAX-strlen(filename)); return fopen(filename,type); } else { return NULL; } } -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 22:19:51 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 12 Message-ID: <74mt27$c4c$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <913239843.29831.1.nnrp-05.d4e4c6f2@news.demon.co.uk> <74msvl$c2k$1@agate.berkeley.edu> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74msvl$c2k$1@agate.berkeley.edu>, Eric J. Korpela wrote: > strncat(filename,"file.typ",_POSIX_PATH_MAX-strlen(filename)); Oops, that "file.typ" should be file. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: pm215@watchdragon.demon.co.uk (Peter Maydell) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 22:20:41 -0000 Organization: dragon cluster Message-ID: <74mt3p$1gd$1@watchdragon.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> Reply-To: pmaydell@chiark.greenend.org.uk NNTP-Posting-Host: watchdragon.demon.co.uk X-NNTP-Posting-Host: watchdragon.demon.co.uk:158.152.121.201 X-Trace: news.demon.co.uk 913262690 nnrp-01:15360 NO-IDENT watchdragon.demon.co.uk:158.152.121.201 X-Complaints-To: abuse@demon.net Lines: 13 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!dispose.news.demon.net!demon!news.demon.co.uk!demon!watchdragon.demon.co.uk!not-for-mail In article <74jjnt$i01$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: >Please point me to a system which is branded UNIX and does not contain >a Bourne shell. If you can't please take back your statement. Fifth Edition UNIX doesn't have the Bourne shell. IRMTI. It has a /bin/sh, but it isn't anything like the Bourne shell. I don't think the Bourne shell came in until Seventh Edition, but I could be wrong there (might have been in 6th Ed.) Peter Maydell [who hopes all the UNIX vendors have rewritten their Bourne shells by now :->] ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:25:23 -0600 Organization: MindSpring Enterprises Lines: 11 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> <74kasb$47f$1@mathserv.mps.ohio-state.edu> <366df034.14353548@news.xmission.com> NNTP-Posting-Host: d1.56.d0.6a X-Server-Date: 10 Dec 1998 04:22:05 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <366df034.14353548@news.xmission.com>, skb@xmission.removethis.com (Scott Brown) wrote: >Unfortunately, there doesn't seem to be a free VMS-alike that will run >on a surplus 80486 PC. Yeah, it's a problem. However, you can get a cheap used VAX, get a Hobbiest license, and load VMS 6.1 on it w/o a charge other than for the media. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:44:06 -0600 Organization: MindSpring Enterprises Lines: 66 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> NNTP-Posting-Host: d1.56.d2.34 X-Server-Date: 10 Dec 1998 04:40:50 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74knnt$8c1@news1.newsguy.com>, mww@microfocus.com (Michael Wojcik) wrote: >I've used an awful lot of OSes, and I've never seen one with an >intuitive UI. Yes, the names of the commonly used Unix commands >are obscure. I didn't find them any more difficult to learn than >those of a dozen other platforms. But this is so thoroughly >subjective an issue that it's hardly worth mentioning, let alone >discussing. Granted. Still, when I had to get an economist and an accountant used to UNIX (which isn't yet a completed project), I had -lots- of fun explaining certain things, like the chmod protection mask. YOU try explaining binary (and octal) to an accountant. I loaned them the Tom Lehrer CD with the "New Math" track on it after I tried explaining. I don't think it helped. >Watchpoints are a hardware issue; they have nothing to do with >software. Either there's hardware assist (like watch registers) or >there isn't. (If the hardware assist is there, but the software >doesn't use it, get a different debugger. They're generally free.) I have to yield this one. I seem to remember that VAX has a tracepoint bit, and watchpoints were implemented in some way I wouldn't have thought of, but that they were both effectively hardware. >Unix has developer tools all over the place. That's mostly what Unix >is - a collection of developer tools (though again that's not a core >OS matter). I humbly submit that you don't know what you're talking >about here. (That may be because Unix is "hard to learn" or some >such, but that's an entirely different matter.) In part, it actually has more to do with the fact that I'm more interested in flaming UNIX in order to vent. I don't like UNIX. I find it much too ad-hoc for my taste. VMS is nice and simple. The main thing I think it's missing as of the last time I looked, was pipes. UNIX pipes are neat. >> UNIX has no calling standard. > >Of course it doesn't. It runs on heterogeneous hardware. Hardware >drives calling standards. Perhaps. However, if variables and such were laid out in a way which was predictable by all language processors and debuggers, I suspect life would be better. This may not be possible. >But your comparisons between VMS and Unix mostly illustrate your >inability or unwillingness to distinguish between hardware and >software issues, or OS and applications, or core OS and third-party >add-ons, or design and implementation, or hardware models and >platforms. When people are posting religious war flamebait on >alt.folklore.computers, they ought to at least strive for a bit of >precision. The problem, from my point of view, is that with VAX/VMS, the two were very tightly coupled. For instance, they put some instructions in hardware even though they wouldn't be used more than once by ANYONE because these things had to be FAST (LDPCTX/SVPCTX), for instance. Others were in software, such as H_FLOATing instructions and some character string instructions, which were not implemented on a particular version of a VAX. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:51:13 -0600 Organization: MindSpring Enterprises Lines: 17 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk> NNTP-Posting-Host: d1.56.d2.34 X-Server-Date: 10 Dec 1998 04:47:55 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!howland.erols.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk>, ejb@klamath.demon.co.uk (Edward John M. Brocklesby) wrote: >We're porting our IRC bot (Ilona) to VMS at some point. (It might well be >me who'll have to do this). But I can't find any documentation for it.. >The system I'll be using has MULTINET TCP/IP installed, but I can't find >anything like section 3 on Unix .. I suppose I just haven't looked hard >enough :) Are there any manuals available online? No, neither are they available on UNIX. Far as I can tell, the only paper manuals available for UNIX are compilations of MAN pages. There are no actual manuals. For VMS, there is help on-line, and more recently they've been made available electronically, but no. And note that VMS is being supplanted by Windows NT too. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 09 Dec 1998 22:51:59 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> NNTP-Posting-Host: d1.56.d2.34 X-Server-Date: 10 Dec 1998 04:48:40 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74mabs$rrl$1@agate.berkeley.edu>, korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >Strange. Are you telling me that VMS directories weren't hacked onto a >flat file system? Can't say. They don't -look- like it. -- Howard S Shubs The Denim Adept ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 9 Dec 1998 23:25:31 GMT Organization: Daedalus Consulting Lines: 79 Message-ID: <74n0tb$i8i$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <366d8031.0@news2.uswest.net> <367591ad.19390965@Rockyd> <366dabe5.0@news3.uswest.net> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Chris Mikkelson wrote: >In article <367591ad.19390965@Rockyd>, >Since each disk/FS block can only contain data from one file, a >file will *use* more disk space than its size would reflect. Thus, >df reports the real used and free space on the disk, rather than >just a total of the file sizes. > >Can you do that on VMS? Not being snide, just asking. DIRECTORY can tell you exactly how much disk space a file takes -- the /SIZE qualifier can be /SIZE=USED, /SIZE=ALLOCATION or /SIZE=ALL. ALLOCATION gives you the actual blocks used, while USED gives the blocks actually holding data. VMS makes no attempt to hide the "overhead", in fact every block of the file system is mapped to a file, including the home block and backup home blocks, file headers, storage bitmap, boot blocks, bad blocks and so-on. Thus, the grand total of DIR/GRAND/SIZE=ALLOC [*...] plus the free space as reported by SHOW DEVICE should[1] exactly match the size of the disk. That said, what directory *doesn't* do is give you the actual byte count for a given file. This is mainly because the record oriented nature of VMS (via RMS) means that a pure byte count is a lot less meaningful than it is in unix where it's normal to have every byte of the file available to the application, whereas VMS applications usually just read the record data. The following bit of DCL will give the byte count of the file (including RMS record lengths, pad bytes and so-forth): $ eof = f$file_attributes(p1, "EOF") ! Block containing EOF $ ffb = f$file_attributes(p1, "FFB") ! First free byte in EOF block $ bytes = (eof - (ffb .ne. 0)) * 512 + ffb (The equivalent can be done in native applications by using the RMS $DISPLAY and grabbing the relevant fields returned.) Note that what's happening here is that these numbers come from the underlying filesystem, not RMS (although RMS is digging them out). RMS applies its own idea of what's in the files on top of the raw length returned from the filesystem. You can tell RMS to step aside and let you have at the filesystem directly, either by constructing calls to the filesystem ACP (or XQP for ODS2 -- an implimentation detail, the interface is the same) yourself, or using the RMS block read/write routines. But this stuff deals with files as arrays of blocks (rather than arrays of characters a la unix), and while it's not hard to layer your own data format over that, there's usually little point, and it's usually easier to create files in Stream_LF (which makes the file, and its representation on disk, look *remarkably* like that normally used by unix applications and utilities) format or Undefined format, where record lengths are defined as the size of your buffer and record positions based on where the last "record" stopped. [1] but often doesn't. The main reason is that files can be deleted while still open, in which case they're removed from the directory but stay active with a valid file header marked for delete, and actually get deleted when the file is closed. If the file isn't closed properly (eg system failure), they stick around and have to be mopped up by ANALYZE/DISK/REPAIR. (Incidentally, I once rescued a system where the system disk had the entire track 0 wiped by a hardware misconfiguration, without restoring from backup (backup?) -- booted off another disk, ANAYLYZE/DISK/REPAIR, then a bit of juggling to get the newly re-created root directory into the right place.) >The equivalent of DIR /TOTAL (which doesn't list the entries, >right?) would be du (disk usage), which will cross mount points. >I can say this is equivalent to the VMS version, assuming that the >VMS "DIR /TOTAL" would cross mount points if VMS had mount points. du can be told not to cross mount points, to follow symlinks or not etc. >Oh, and why did you "painstakingly" add all the file sizes when >you could use awk? Yep. This really is where unix does have its charm -- utilities are generally designed to be plumbed together in pipelines. While you can do this with VMS you tend to end up having to parse extraneous stuff, and it's not generally pretty. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: deke.spamblock@generous.net Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 00:14:51 GMT Organization: GenerousCity is a virtue - find romance at http://generous.net Lines: 62 Message-ID: <366f0671.114112447@news.bright.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: paul-cas1-cs-20.dial.bright.net X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newspeer1.nac.net!netnews.com!howland.erols.net!newspump.sol.net!nntp.msen.com!news.altair.com!NewsNG.Chicago.Qual.Net!news.bright.net!not-for-mail On 9 Dec 1998 16:34:55 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >In article <74kofa$8tf@news1.newsguy.com>, >Michael Wojcik wrote: >>The shell, Bourne or otherwise, isn't part of "straight Unix", if >>by that we mean the core OS. >The shell is part of POSIX.2 (the second and third letters stand for >"operating system") and also part of the OpenGroup Unix specification. >If you delete /bin/sh what you have left will not be Unix (since it >will not pass the POSIX or OpenGroup specifications). What was it that AT&T was calling unix from 1970 until 1988? POSIX and OpenGroup didn't exist at that time. >>The shell has no special ties to the core OS, regardless of whether it >>was shipped with the OS or added later. The original poster's >>statement is perfectly valid. >It's -part- of the OS. Unless you don't understand the difference >the kernel and the OS. Well, *I* understand the difference between the kernel and the OS. The operating system is built like an onion. There is the kernel on the inside. The system calls and the device drivers are the next layer of the onion. The file system, networking, print queue, and scheduling systems are the outer layer. Application programs, utilities, and shells are *not* part of the onion. If you would like a diagram to look at, check the SCO documentation for Unix. >I am signficantly aware of this. At work I'm running a bare executable >on top of the Linux kernel (as an embedded system)-- no shell. This >has absolutely NOTHING to do with Unix though -- it certainly would not >be branded Unix by the OpenGroup or be POSIX compliant (and not because >it's Linux, but because it doesn't have a shell and other core >utilities). Sure, and if you put a firebird decal on a Mustang, GM would probably let you call it a Pontiac. Sure. Right. >A massive quantity of off-the-shelf application which effectively say >"requires (e.g.) Solaris to run" require not just the kernel, but lots >of helper programs (of which the most common is probably the shell). >Please note that they say "requires Solaris" not requires "Solaris plus >/bin/sh and /bin/ls and /bin/cp", etc. Therefore one can conclude >that the term "Solaris" means not the kernel but the kernel as well >as included utilities. It's like this for any Unix. But what does that have to do with anything? Solaris is not "straight unix" - it has a number of features that other unixes do not have. That's why a program will say "requires Solaris to run" instead of "requires any Unix operating system." ------------------------ Let love find you! http://generous.net A list for flirting GenerousSingles-subscribe@onelist.com Over The Hill Gang GenerousSinglesOTHG-subscribe@onelist.com Personal Ads list GenerousProfiles-subscribe@onelist.com ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 00:57:30 GMT Organization: University of Illinois at Urbana-Champaign Lines: 35 Message-ID: <74n69q$3n4$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <366f0671.114112447@news.bright.net> NNTP-Posting-Host: sparc43.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!128.174.5.49!vixen.cso.uiuc.edu!tsmurphy In article <366f0671.114112447@news.bright.net>, wrote: >Well, *I* understand the difference between the kernel and the OS. Everything you describe below is part of the kernel (well, except the print queue) in a normal Unix system. What you describe is the structure of the kernel not the OS. >The operating system is built like an onion. There is the kernel on the inside. >The system calls and the device drivers are the next layer of the onion. The >file system, networking, print queue, and scheduling systems are the outer >layer. Application programs, utilities, and shells are *not* part of the onion. >If you would like a diagram to look at, check the SCO documentation for Unix. Onion THIS: the cover of "The Design of the Unix Operating System" by Maurice J Bach. The shell and core utilties are in the third inner most layer of the system. I'd say MJB is a little more authoritative than the SCO documentation, even in the Unix community. >But what does that have to do with anything? Solaris is not "straight >unix" - it has a number of features that other unixes do not have. >That's why a program will say "requires Solaris to run" instead of >"requires any Unix operating system." You're totally missing the point. I'm talking about basic features such as #!/bin/sh script. I have never seen a #!/bin/sh script that said "runs on Unix systems with /bin/sh installed". Rather they say "runs on Unix system" because /bin/sh is part of Unix. -- Terry ###### From: william.hamblen@nashville.com (William Hamblen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 02:45:30 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 8 Message-ID: <36762d57.170961415@news.nashville.com> References: <366df4be.44034423@news.bright.net> <1571.647T270T6763223@sky.bus.com> NNTP-Posting-Host: 207.65.139.220 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: 913257553 4TJCV727N8BDCCF41C usenet54.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Agent 1.5/32.452 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!nntp.news.xara.net!xara.net!SupernewsUK!supernews.com!SnNA!Supernews69!not-for-mail On 09 Dec 98 11:16:05 -0800, "Charlie Gibbs" wrote: > ... MS-DOS's >COMMAND.COM, on the other hand, is "the command processor" >simply because no replacement has really caught on. There does at least exist 4DOS, and probably others. ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 06:00:17 GMT Organization: La Trobe University Lines: 17 Message-ID: <74no1h$19v$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 913269617 1343 131.172.12.11 (10 Dec 1998 06:00:17 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 10 Dec 1998 06:00:17 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd jmfbahciv@aol.com wrote: I wrote: : >>This must have been around 1983. I guess other operating systems have : >>a fair way to go.... : One of my frustrations is seeing our wheels reinvented as if they : had never existed. I guess I should have written "I guess other _modern_ operating systems...". Take NT for example (please :-) -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 06:46:22 GMT Organization: La Trobe University Lines: 19 Message-ID: <74nqnu$1si$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 913272382 1938 131.172.12.11 (10 Dec 1998 06:46:22 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 10 Dec 1998 06:46:22 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!dca1-hub1.news.digex.net!digex!cpk-news-hub1.bbnplanet.com!su-news-feed4.bbnplanet.com!news.gtei.net!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Eric J. Korpela (korpela@islay.ssl.berkeley.edu) wrote: : My method of coping was to cease using VMS. Is there anyone out there : who didn't alias "SET DEFAULT" to something else? "CD" maybe? Well I use the alias cp for SET DEFAULT. Of course, this doesn't make sense unless you were a TOPS-10 user and remember SETSRC :-) : Well, it wouldn't be that difficult to create the equivalent of logical : names in a unix shell. It's just not really necessary. Why create something : like that when a symbolic link will work. You can always do this with VMS too. See SET FILE/ENTER. -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 06:55:25 GMT Organization: La Trobe University Lines: 26 Message-ID: <74nr8t$1si$2@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 913272925 1938 131.172.12.11 (10 Dec 1998 06:55:25 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 10 Dec 1998 06:55:25 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!news.mel.connect.com.au!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Robert Billing (unclebob@tnglwood.demon.co.uk) wrote: : Howard S Shubs wrote: : > Logical names are, in part, a way for system managers to alias things in : > order to organize the system better. As a UNIX-head, you're familiar with : > the concept of the "~" directory. Which is better, using that notation or : > using the full path? : The other virtue of logical names is that they can be different for : different users. It is perfectly possible to set up the system in such a : way that the system login file creates logical names for each session : which depend on the user ID, time, phase of the moon, number of : [mongooses|mongeese] in stock... This property is vital if you run a mixed architecture VMS cluster where you need to setup logicals pointing to the correct architecture executable directory depending on the node being employed. If you really want to talk about what differentiates VMS from basically every other commercially available operating system, talk clusters. -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 10:26:05 GMT Organization: Daedalus Consulting Lines: 25 Message-ID: <74o7jt$6d6$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article , Howard S Shubs wrote: >I have to yield this one. I seem to remember that VAX has a tracepoint >bit, and watchpoints were implemented in some way I wouldn't have thought >of, but that they were both effectively hardware. It's done with page protections -- the page containing the watchpoint is marked read-only, and if the code attempts to modify it, it causes an exception which is handled by the debugger. The debugger then takes a copy of the page, makes the page writable, sets the trace bit and dismisses the exception. The instruction doing the write completes, and then triggers a trace exception. The debugger catches that, marks the page read-only again, compares the modified page, and if the modification occurred on a location that it was watching it takes the appropriate action, otherwise it allows the code to continue. I might be a little off on fine detail, but you get the idea. That mechanism can be done on pretty much anything with paged memory management and OS services permitting the debugger to futz with the page protections. (Note that under VMS, the guts of the debugger run in the context of the process running the code being debugged.) -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 10:29:35 +0000 Organization: Tanglewood Message-ID: <366FA28F.40D8C49E@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> <74o7jt$6d6$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913292805 nnrp-09:23306 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 13 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!woodstock.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Don Stokes wrote: > That mechanism can be done on pretty much anything with paged memory > management and OS services permitting the debugger to futz with the > page protections. (Note that under VMS, the guts of the debugger run Does this mean it could be done by Linux? I suspect it could... -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 11:04:05 GMT Organization: Daedalus Consulting Lines: 24 Message-ID: <74o9r5$r28$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!feeder.qis.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74mabs$rrl$1@agate.berkeley.edu>, Eric J. Korpela wrote: >Strange. Are you telling me that VMS directories weren't hacked onto a >flat file system? I wrote a long ramble about RSX and its directory structure (since the RSX Files-11 ODS-1 filesystem is the direct ancestor of the VMS ODS-2 structure, and it used [p,pn] directories), but then I thought to go and look at the ODS-1 and ODS-2 specs. Both contain the following paragraph: 4.1 Directory Hierarchies Since directories are files with no special attributes, directories may list files that are in turn directories. Thus, the user may construct directory hierarchies of arbitrary depth and complexity to structure his files as he pleases. The ODS-1 spec I have is dated 1975, well before VMS. So I think we can safely dispose of the idea that VMS's hierarchical directories were "hacked onto a flat file system". -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 11:23:03 GMT Organization: Daedalus Consulting Lines: 34 Message-ID: <74oauo$5ia$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <366dabe5.0@news3.uswest.net> <74n0tb$i8i$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article , Howard S Shubs wrote: >In article <74n0tb$i8i$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz >(Don Stokes) wrote: >>blocks, bad blocks and so-on. Thus, the grand total of >>DIR/GRAND/SIZE=ALLOC [*...] plus the free space as reported by SHOW DEVICE >>should[1] exactly match the size of the disk. > >Unless things have changed, this is not quite true. DIR/SIZ=ALL tells you >the number of disk blocks the file takes, yes. However, if you've got >disk quotas enabled, you'll find that adding up all the allocated blocks >will not jibe with the number given by SHOW QUOTA. This is because not >only does the quota charge the file blocks against you, but also the file >header blocks for those files you own. Each file as at least one file >header block. Some have more, for various reasons. Disk quotas do "charge" you for the header block(s), true. But these blocks are mapped in the index file [000000]INDEXF.SYS. So the quota facility charges you that part of INDEXF.SYS that you've "taken". There's no otherwise unaccounted for space involved. What I wrote (if you include the footnote) is absolutely true for total disk usage. INDEXF.SYS is a bit of a dumping ground for filesystem structures. It maps (in order) the boot block, home block, backup copies of the home block, copies of the the file header for INDEXF.SYS itself, bitmap of available file headers, then one block each for each file header (including file headers for INDEXF.SYS, the MFD, quota file and so-forth). Note that the home block and copies of the INDEXF.SYS are in elsewhere-defined places; INDEXF.SYS just maps them -- obviously, the file system can't find INDEXF.SYS from a file header located inside INDEXF.SYS... -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 98 12:35:46 GMT Organization: UltraNet Communications, Inc. Lines: 28 Message-ID: <74ogik$nrp$1@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <74no1h$19v$1@news.latrobe.edu.au> NNTP-Posting-Host: d3.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 10 Dec 1998 12:59:00 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newspeer.monmouth.com!news.ultranet.com!d3 In article <74no1h$19v$1@news.latrobe.edu.au>, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >jmfbahciv@aol.com wrote: > >I wrote: > >: >>This must have been around 1983. I guess other operating systems have >: >>a fair way to go.... > >: One of my frustrations is seeing our wheels reinvented as if they >: had never existed. > >I guess I should have written "I guess other _modern_ >operating systems...". Sorry, I just had to vent (as you would understand since I remember your name on some SPRs). >Take NT for example (please :-) You may keep it. I didn't touch VMS (unless forced to) because of being offended. NT is far, far more advanced in that category. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 98 12:39:09 GMT Organization: UltraNet Communications, Inc. Lines: 22 Message-ID: <74ogov$nrp$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> NNTP-Posting-Host: d3.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 10 Dec 1998 13:02:23 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed.wirehub.nl!news.maxwell.syr.edu!newsfeed.xcom.net!news.ultranet.com!d3 In article <74nqnu$1si$1@news.latrobe.edu.au>, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >Eric J. Korpela (korpela@islay.ssl.berkeley.edu) wrote: > >: My method of coping was to cease using VMS. Is there anyone out there >: who didn't alias "SET DEFAULT" to something else? "CD" maybe? > >Well I use the alias cp for SET DEFAULT. Of course, this doesn't make >sense unless you were a TOPS-10 user and remember SETSRC :-) Ahem....or PATH? Have you been finding it amusing the way people seem to be confusing paths, file structure heirarchy, logical names? I've been trying to figure out whether it's due to simple confusion or some OS really implemented the mess. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 98 12:44:19 GMT Organization: UltraNet Communications, Inc. Lines: 29 Message-ID: <74oh2l$nrp$3@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: d3.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 10 Dec 1998 13:07:33 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d3 In article <74m8rf$5ev$1@vixen.cso.uiuc.edu>, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >In article <74kofa$8tf@news1.newsguy.com>, >Michael Wojcik wrote: > >>The shell, Bourne or otherwise, isn't part of "straight Unix", if >>by that we mean the core OS. > >The shell is part of POSIX.2 (the second and third letters stand for >"operating system") and also part of the OpenGroup Unix specification. > >If you delete /bin/sh what you have left will not be Unix (since it >will not pass the POSIX or OpenGroup specifications). > >>The shell has no special ties to the core OS, regardless of whether it >>was shipped with the OS or added later. The original poster's >>statement is perfectly valid. > >It's -part- of the OS. Unless you don't understand the difference >the kernel and the OS. You should rephrase that to be "don't understand the difference between the kernal and the OS _of UNIX_"; there were other OSes that didn't have quite the same differentiation. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 98 13:02:30 GMT Organization: UltraNet Communications, Inc. Lines: 132 Message-ID: <74oi4r$nrp$4@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> NNTP-Posting-Host: d3.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 10 Dec 1998 13:25:47 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d3 In article <74kkr4$6t9@news1.newsguy.com>, mww@microfocus.com (Michael Wojcik) wrote: > >In article <74glh7$j1n$1@strato.ultra.net>, jmfbahciv@aol.com writes: >> [somebody's hacked off the atributes and it tweren't me] >> >> In article <366b39eb.3006571501@news.cyberport.com>, >> tangentSPAMCATCHER@cyberport.com (Warren Young) wrote: > >> >I've only been working with computers seriously since about 1990 (and >> >a nerdy little whiz kid since 1985), but it seems to me that you could >> >do this with a preemptible, multitasking kernel. No, not a kernel >> >that provides multitasking to the user programs, but that can be >> >preempted itself. This is probably easiest to do with a >> >microkernel-style implementation. I know NT has a semi-micro-kernel >> >setup, and I suspect that VMS does, too. >> >> You must consider the case where your microkernel preempts itself. > >The AIX kernel, as of v3, is preemptible as well as pageable. And >it's not even a microkernel. It's not a trivial design, but it's >not prohibitively difficult. Mostly it means putting a lot of >fast synchronization mechanisms in at the right places. But, if I understand the implementation, is it re-entrant? > >> >So, when the swapper or pager process decides it needs to move a block >> >of memory to the swap file to free up some memory, it simply calls the >> >disk I/O routines to do this. They then realize that the volume in >> >question is apparently unavailable. Now, since it hasn't been >> >formally dismounted, it's probably a temporary error, so the call >> >blocks. The pager then sits there, blocking on a disk I/O call until >> >it completes. > >AIX, like some other modern Unix implementations, doesn't have a >separate pager process. All disk I/O (except I/O to special >partitions that are reserved to special third-party applications, >like DBMSes, and largely ignored by the OS) goes through the kernel's >virtual memory manager. That means all swap and file I/O is >memory-mapped, under the covers ("covers" being the read, write, etc. >system calls) if necessary. OK, so I'm assuming that this VM manager is a bottleneck. And that doesn't sound very good for performance (understand, that I come from a group who was terribly fussy about good performance w.r.t the user). It also means that general I/O is funneled through the VM code. That just gives me the shivers--of course, it could be that I'm biased to the hilt :-). > >The VMM, which is a kernel lightweight process, won't block forever >on I/O problems because it has a watchdog timer. Good gads, it shouldn't! The requestor should probably have the timeout for retries or whatever is appropriate for that particular device medium (note that this should be at the device driver level). > I haven't looked >into how the kernel behaves in these cases, but I imagine its action >is based on the severity of the problem (probably transient? probably >permanent? affecting user code? a critical kernel process?). I don't know since it seems that the philosophy of the implementation seems upsidedown to me :-). > >> Assuming it is a temporary error (whatever that means...there are >> myriads of categories), do you tell someone about the error? And, >> if you do think someone should be notified, who? > >In Unix, that's a configuration issue, if the problem is non-fatal; >it's logged via the syslog facility or something similar in most >flavors, and you have various options for what "logging" involves. But what piece defines fatal? > >If the kernel panics, that's a different story, but again details >are implementation-specific. You just confused me. Which implementation? Site, user, OS, hardware? >Typically the kernel dumps core to >the save device, if it can, and shuts down or reboots. Many modern >machines have some facility for storing some error information in >non-volatile memory, too. TOPS10 also had a "dump core and continue" stopcode category. That was the basis for implementing the "suspend" option of Symmetric Multi-Processing. > >> >Alternately, that disk I/O call could complete after passing the data >> >off to the cache subsystem. The cache subsystem would then wait >> >patiently for the disk to become available again. >> >> Thus you leave the user or program with the understanding that the >> data has been written in iron. Now the next action by the same >> user (or, more likely, another user) is to read back the newly >> written data. How do you ensure data integrity? > >With the all-I/O-via-VMM architecture, it's not a problem. The OS >knows which pages are in RAM and which are not. It is a problem if you wanted to do symmetric multi-processing. It doesn't seem to allow I/O that is synchronous with the user requests but uses an asynchonous process. (Or, I'm simply not understanding what you're describing.) > As you pointed out >in another note in this stream, there are performance implications, >but actually the AIX architecture performs very well for the situation >you describe, given a reasonable amount of RAM. That's partly because >it has hardware assist in the form of reverse translation lookaside >buffers. The OS does provide a system call to force immediate >scheduling of a dirty page commit, for applications that want a little >more assurance; but of course many drive/controller combinations these >days will take queued requests and pretend they're done before they've >actually been committed to physical media. > >But as you say that's not an ideal architecture for all applications, >which is why you have DBMSes that write to raw partitions and don't >go through the OS' VMM. And that sentence puts shivers of dread all up and down my spine. This was an anathema in TOPS10 :-). /BAH Subtract a hundred and four for e-mail. ###### From: Mike Swaim Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> Organization: PointeCom User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/2.2.8-RELEASE (i386)) Lines: 26 Message-ID: Date: Thu, 10 Dec 1998 15:06:25 GMT NNTP-Posting-Host: 209.127.0.130 X-Trace: news2.giganews.com 913302385 209.127.0.130 (Thu, 10 Dec 1998 09:06:25 CDT) NNTP-Posting-Date: Thu, 10 Dec 1998 09:06:25 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-a.ais.net!ais.net!nntp.giganews.com!news2.giganews.com.POSTED!gemini.c-com.net!swaim jmfbahciv@aol.com wrote: : In article <74kkr4$6t9@news1.newsguy.com>, : mww@microfocus.com (Michael Wojcik) wrote: :>But as you say that's not an ideal architecture for all applications, :>which is why you have DBMSes that write to raw partitions and don't :>go through the OS' VMM. : And that sentence puts shivers of dread all up and down my spine. : This was an anathema in TOPS10 :-). I believe that MS SQL Server through version 6.5 does something similar, although it doesn't use its own partitions. (I think that it just bypasses the higher levels of the file system.) I've been told that the reason is that the disk reads/writes are sufficently random from an external processes point of view that the tricks used by the OS to speed things up (notably read ahead) have the opposite effect. I've also heard of a LISP interpreter that used memory locations to encode type information. It would occasionally tell the memory manager that it was going to do essentially random reads and writes, and not to bother with its normal page replacement algorithm. (I don't remember the system, though.) -- Mike Swaim, Avatar of Chaos: Disclaimer:I sometimes lie. Home: swaim@c-com.net Alum: swaim@alumni.rice.edu Quote: "Boingie"^4 Y,W&D ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 17:25:33 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 18 Message-ID: <74p06d$oso$1@agate.berkeley.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74nr8t$1si$2@news.latrobe.edu.au> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <74nr8t$1si$2@news.latrobe.edu.au>, Huw Davies wrote: >If you really want to talk about what differentiates VMS from basically >every other commercially available operating system, talk clusters. What exactly do clusters give you that NFS+(YP || NIS) doesn't give you? An artificially short maximum path between computers? (Maybe this has changed since I used VMS, but machines in a cluster needed to be basically on the same subnet.) At one point I worked on the (physically) largest VAX cluster in the world, which had a 25km optical fiber link between two nodes. That was a long, long time ago, though.) Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 18:52:44 +0100 Organization: [Posted via] Interactive Networx Lines: 10 Message-ID: <1djto6g.swejxkh3o4o4N@n241-84.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74mmn1$4l7$3@irk.zetnet.co.uk> NNTP-Posting-Host: n241-84.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!fu-berlin.de!cs.tu-berlin.de!unlisys!news.snafu.de!jnickelsen wrote: > Some people use GNU Emacs as their shell. For fun I tried that once, but it didn't work. I think Emacs tried to execute commands using my login shell (i. e. Emacs) or something like that, but it's too long ago and I don't remember what exactly failed. -- Juergen Nickelsen ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 18:52:47 +0100 Organization: [Posted via] Interactive Networx Lines: 33 Message-ID: <1djtoy1.a1f9ypmi4mj7N@n241-84.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> NNTP-Posting-Host: n241-84.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!wuff.mayn.de!unlisys!news.snafu.de!jnickelsen Huw Davies wrote: > Eric J. Korpela (korpela@islay.ssl.berkeley.edu) wrote: [...] > : Why create something like that when a symbolic link will work. > > You can always do this with VMS too. See SET FILE/ENTER. I have heard this is more like Unix's hard link -- what is it? For those not familiar with the Unix concepts, I summarize the properties of both: hard links - are directory entries of files; each file in a directory has at least one - the original directory entry of a file and additional ones are not of a different kind - you can delete the original entry and the file is still accessible through the other entry/entries - a hard link cannot refer to a file on another file system symbolic links - are just special files containing a pathname (but is, by the OS, interpreted on open() etc. as the file it refers to) - may point to a non-existent pathname, and do so if the file or directory they point to is deleted - may point to a file/directory on another file system (corrections and additions welcome) -- Juergen Nickelsen ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 18:52:50 +0100 Organization: [Posted via] Interactive Networx Lines: 15 Message-ID: <1djtpzr.4qdlk71n5ea98N@n241-84.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <74mmn9$4l7$5@irk.zetnet.co.uk> NNTP-Posting-Host: n241-84.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!news-fra.maz.net!newsfeed.ecrc.net!unlisys!news.snafu.de!jnickelsen wrote: > On the other hand, since a directory is merely a list of filenames and > i-nodes, you can open a directory as a file, which eases the pain > somewhat. Not exactly -- NFS mounted filesystems don't allow opening a directory as a file, since the host the file system resides on might not be running Unix (and therefore not even have directory files, like the Mac). While you may open directories of local file systems as files, it is more convenient to use the opendir(), readdir(), etc. calls in either case. -- Juergen Nickelsen ###### Date: 10 Dec 98 10:10:32 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> Message-ID: <1456.648T1620T6105137@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 34 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 10 Dec 1998 13:21:58 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.104 In article hshubs@mindspring.com (Howard S Shubs) writes: >Granted. Still, when I had to get an economist and an accountant used >to UNIX (which isn't yet a completed project), I had -lots- of fun >explaining certain things, like the chmod protection mask. YOU try >explaining binary (and octal) to an accountant. I loaned them the >Tom Lehrer CD with the "New Math" track on it after I tried explaining. >I don't think it helped. I doubt it would. As Tom Lehrer himself said, It's so simple, so very simple That only a child can do it. I loved the way the back of the album cover showed the arithmetic problem that he was solving in the song - in both base 10 and base 8. ("Base 8 is just like base 10, really - if you're missing two fingers.") Speaking of Tom Lehrer, the lyrics and music to many of his songs are available in the book "Too Many Songs by Tom Lehrer (and not enough drawings by Ronald Searle)". It includes that delightful Gilbert & Sullivan take-off in which he names every element in the periodic table. At least every one that was known at the time: These are the only ones of which the news has come to Hah-vahd, And there may be many others but they haven't been discah-vahd. (He always was good at twisting a rhyme out of just about anything.) -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: Peter Ludemann Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 10:50:46 -0800 Organization: Inxight Software, Inc. http://www.inxight.com Lines: 22 Sender: ludemann@wistaria Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> <74o7jt$6d6$1@news.wlg.netlink.net.nz> <366FA28F.40D8C49E@tnglwood.demon.co.uk> NNTP-Posting-Host: 13.242.52.14 X-Newsreader: Gnus v5.5/Emacs 20.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.corridex.com!su-news-hub1.bbnplanet.com!su-news-feed1.bbnplanet.com!news.gtei.net!parc!womble.xsoft!usenet Robert Billing writes: > Don Stokes wrote: > > > That mechanism can be done on pretty much anything with paged memory > > management and OS services permitting the debugger to futz with the > > page protections. (Note that under VMS, the guts of the debugger run > > Does this mean it could be done by Linux? I suspect it could... Not only can it be done; it is done. Take a look at the ElectricFence memory-allocation debugging package which uses this technique. And, BTW, for all of you who lament Unix's expansion of file names in the shell (hmmm, why hasn't anybody mentioned DataGeneral's OS which also had the program expand filenames, just like VMS?) ... take a look at POSIX's fnmatch() and glob() functions. You just have to quote the file names in the command line and let the program do its stuff. [tar has been doing this for a long time.] - peter ludemann ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> X-Newsreader: trn 4.0-test67 (15 July 1998) From: cmikk@tig.oss.uswest.net (Chris Mikkelson) NNTP-Posting-Host: 204.147.85.37 Message-ID: <36702c1e.0@news2.uswest.net> Date: 10 Dec 1998 14:16:30 +0600 X-Trace: 10 Dec 1998 14:16:30 +0600, 204.147.85.37 Lines: 15 X-Report: Report abuse to abuse@uswest.net. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsfeed.wli.net!su-news-hub1.bbnplanet.com!su-news-feed4.bbnplanet.com!news.gtei.net!newsfeed1.uswest.net!news1.uswest.net!not-for-mail In article <366E208A.E2CCAB80@tnglwood.demon.co.uk>, Robert Billing wrote: > The other virtue of logical names is that they can be different for >different users. It is perfectly possible to set up the system in such a >way that the system login file creates logical names for each session >which depend on the user ID, time, phase of the moon, number of >[mongooses|mongeese] in stock... Actually, you can do something similar in Unix, too. The closest Unix has to a "system login" are the /etc/csh.login (for csh) and /etc/profile (for sh) files. In there, you can run programs to determine user ID, time, phase of moon, mongoose counting, etc. and use their results to set environment variables. -Chris ###### From: "Jack Peacock" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 15:55:11 -0800 Organization: Simco Lines: 13 Message-ID: <74pmuv$470$1@remarQ.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> <74p9fu$a0n@news3.newsguy.com> NNTP-Posting-Host: 207.168.124.145 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Trace: 913334047 TWNY2XU9G7C91CFA8C usenet53.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!feeder.qis.net!newsfeed-east.supernews.com!supernews.com!Supernews69!not-for-mail Michael Wojcik wrote in message <74p9fu$a0n@news3.newsguy.com>... >Note that interrupt handlers, while preemptible (by kernel processes >at higher priority), are not pageable and presumably not reentrant. > Why shouldn't interrupt handlers be reentrant? If you have multiple instances of the same peripheral but different interrupt vectors for each controller, why not use the same interrupt handler code but separate data? Jack Peacock ###### From: "Jack Peacock" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 16:08:42 -0800 Organization: Simco Lines: 19 Message-ID: <74pnt2$c1v$1@remarQ.com> References: <73ela0$51h@hermes.acs.unt.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74nr8t$1si$2@news.latrobe.edu.au> <74p06d$oso$1@agate.berkeley.edu> <74pmah$bsc$1@news.latrobe.edu.au> NNTP-Posting-Host: 207.168.124.145 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Trace: 913335010 TWNY2XU9G7C91CFA8C usenet57.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!newsfeed-east.supernews.com!supernews.com!Supernews69!not-for-mail Huw Davies wrote in message <74pmah$bsc$1@news.latrobe.edu.au>... >Eric J. Korpela (korpela@islay.ssl.berkeley.edu) wrote: > >: What exactly do clusters give you that NFS+(YP || NIS) doesn't give you? > >Reliability. > Hmm, what's the equivalent of the Cluster Lock Manager on Unix? Or queue failover, where the queue manager automatically restarts a queue on a backup node if the first node fails? Can you redirect cluster traffic on Unix machines over private net backbones? What's the Unix equivalent of an HSC? Can you group a whole Unix cluster to answer to a single disk server name (for PC-NFS) like in Pathworks? Jack Peacock ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 18:41:04 GMT Organization: Laser-Scan Ltd. Lines: 17 Message-ID: <74p4k0$18u@relay.lsl.co.uk> References: <74drv4$rah$2@ligarius.ultra.net> <74hddu$ukl$3@irk.zetnet.co.uk> <74mff9$msa@relay.lsl.co.uk> <74mmvu$1lf@cpca3.uea.ac.uk> NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!diablo.theplanet.net!join.news.pipex.net!pipex!news.lsl.co.uk!benh Andrew Cadley (a.p.cadley@uea.ac.uk) wrote: : Ben Hutchings wrote in message <74mff9$msa@relay.lsl.co.uk>... : :Yes, the PC BIOS only supports synchronous I/O. That is why no real : :OS for the PC uses the BIOS for I/O once it has booted. : So what about Windows then? ;-) Even Windows, since version 3.11. It can still use BIOS calls if it has no suitable drivers, though. I believe that's what the "32-bit disk access" setting in Windows 3.11 controlled. -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 18:44:38 GMT Organization: Laser-Scan Ltd. Lines: 14 Message-ID: <74p4qm$18u@relay.lsl.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74jv4i$t2a$1@vixen.cso.uiuc.edu> <74k6sf$dr5$1@agate.berkeley.edu> <74luo6$frj$1@news.wlg.netlink.net.nz> <913239843.29831.1.nnrp-05.d4e4c6f2@news.demon.co.uk> NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed.wirehub.nl!diablo.theplanet.net!join.news.pipex.net!pipex!news.lsl.co.uk!benh Edward John M. Brocklesby (ejb@klamath.demon.co.uk) wrote: : On about 09 Dec 1998 17:04:02 +0200, Ariel Scolnicov might have written: : >system thing (can you run {z,tc,ba,...}sh on anything unUnixy?). This : Yes. AmigaOS. And ~ even works there. The shells don't work very well, though, since they can only use vfork and not fork. I've seen pdksh and know of csh for AmigaOS, but there's no bash port. GNU-Win32 includes what seems to be a completely functional bash port, though. -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 19:22:36 GMT Organization: Micro Focus Inc. Lines: 93 Message-ID: <74p71s$86r@news3.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> NNTP-Posting-Host: p-803.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.uk.ibm.net!ibm.net!logbridge.uoregon.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!mww In article , hshubs@mindspring.com (Howard S Shubs) writes: > In article <74knnt$8c1@news1.newsguy.com>, mww@microfocus.com (Michael > Wojcik) wrote: > > >I've used an awful lot of OSes, and I've never seen one with an > >intuitive UI. [...] > > Granted. Still, when I had to get an economist and an accountant used to > UNIX (which isn't yet a completed project), I had -lots- of fun explaining > certain things, like the chmod protection mask. YOU try explaining binary > (and octal) to an accountant. I loaned them the Tom Lehrer CD with the > "New Math" track on it after I tried explaining. I don't think it helped. I won't disagree with you there. None of the Unix shells I've ever used were conducive to non-programmer use. Ditto most of the standard commands. I'd much rather train a non-programmer on VMS/DCL than, say, Unix and ksh (assuming a week or two to use VMS intensively and recover all those details I've forgotten...) myself. I don't think my Tom Lehrer CD has "New Math". It has "Lobachevsky", though. > >Unix has developer tools all over the place. That's mostly what Unix > >is - a collection of developer tools (though again that's not a core > >OS matter). I humbly submit that you don't know what you're talking > >about here. (That may be because Unix is "hard to learn" or some > >such, but that's an entirely different matter.) > > In part, it actually has more to do with the fact that I'm more interested > in flaming UNIX in order to vent. I don't like UNIX. I find it much too > ad-hoc for my taste. VMS is nice and simple. The main thing I think it's > missing as of the last time I looked, was pipes. UNIX pipes are neat. Yes, I was a bit out of line there. Unix is really targetted toward the kinds of programmers the _New Hacker's Dictionary_ calls "toolsmiths" - people who like to pick up odd things here and there and put them together. It's a game. If you're not in the mood to play it, it's bound to be annoying. > >> UNIX has no calling standard. > > > >Of course it doesn't. It runs on heterogeneous hardware. Hardware > >drives calling standards. > > Perhaps. However, if variables and such were laid out in a way which was > predictable by all language processors and debuggers, I suspect life would > be better. This may not be possible. It's tough to specify that across hardware lines. Support for portability could be better, though. The system headers for things like symbol table entries tend to be a bit opaque. (I added debugging symbol generation to a compiler running under SunOS 4 some years back, and spent a fair bit of time compiling little C fragments to assembly to see what a "real" compiler did.) > >[my somewhat excessive hardware vs software flame snipped] > > The problem, from my point of view, is that with VAX/VMS, the two were > very tightly coupled. For instance, they put some instructions in > hardware even though they wouldn't be used more than once by ANYONE > because these things had to be FAST (LDPCTX/SVPCTX), for instance. Others > were in software, such as H_FLOATing instructions and some character > string instructions, which were not implemented on a particular version of > a VAX. Yes. I'd say the VAX is the second-best CISC architecture I've ever seen, and the best for its time (again, in my experience). I'd rate the AS/400 slightly higher because of its additional levels of abstraction and object orientation (even CISCier than VAX), but it came much later and benefitted from years of advances in hardware engineering, which left a lot of headroom; and even so the first 400s were abysmally slow at the low end. And because, as you say, VMS was tightly coupled to the VAX architecture (though not so tightly that DEC couldn't move it to the Alpha), and that does simplify things like calling conventions and debuggers. I'd say that it was more important, though, that VMS was controlled by DEC, and they could to a large extent enforce particular conventions. That's a standard tradeoff between so-called "open" and "proprietary" systems (though we could debate what those terms really mean until we're all running WinNT). Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University You brung in them two expert birdwatchers ... sayin' it was to keep us from makin' dern fools of ourselfs ... whereas it's the inherent right of all to make dern fools of theirselfs ... it ain't a right held by you official types alone. -- Walt Kelly ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 20:04:14 GMT Organization: Micro Focus Inc. Lines: 217 Message-ID: <74p9fu$a0n@news3.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> NNTP-Posting-Host: p-190.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!194.165.93.117.MISMATCH!newshub.bart.net!newsfeed.wirehub.nl!news-peer.gip.net!news.gsl.net!gip.net!logbridge.uoregon.edu!pln-w!spln!extra.newsguy.com!newsp.newsguy.com!mww In article <74oi4r$nrp$4@ligarius.ultra.net>, jmfbahciv@aol.com writes: > In article <74kkr4$6t9@news1.newsguy.com>, > mww@microfocus.com (Michael Wojcik) wrote: > > > >In article <74glh7$j1n$1@strato.ultra.net>, jmfbahciv@aol.com writes: > >> [somebody's hacked off the atributes and it tweren't me] > >> > >> In article <366b39eb.3006571501@news.cyberport.com>, > >> tangentSPAMCATCHER@cyberport.com (Warren Young) wrote: > > > >> >I've only been working with computers seriously since about 1990 (and > >> >a nerdy little whiz kid since 1985), but it seems to me that you could > >> >do this with a preemptible, multitasking kernel. [...] > >> > >> You must consider the case where your microkernel preempts itself. > > > >The AIX kernel, as of v3, is preemptible as well as pageable. And > >it's not even a microkernel. It's not a trivial design, but it's > >not prohibitively difficult. Mostly it means putting a lot of > >fast synchronization mechanisms in at the right places. > > But, if I understand the implementation, is it re-entrant? Some parts of it are, at least as I understand it. It was designed to support real-time processes (though I'll let others decide to what extent that effort was successful), and so the kernel is preemptible, pageable, reentrant, and supports fixed absolute process priorities. It has fast synchronization mechanisms, since obviously there are a lot of critical sections in a full Unix-style kernel. That's why, for example, the "cs" (compare and swap) atomic system call provided by the AIX 3 kernel used SVC 2 (supervisor call 2) rather than SVC 1, as all other system calls did; SVC 1 calls could preempt one another and reenter. SVC 2 *wasn't* preemptible. (cs was dropped in AIX 4, because it was too difficult to get it working nicely on multiprocessor machines. It was a hack, albeit a relatively neat one.) My main reference for this (besides a bunch of AIX internals discussions I sat in on when I was at IBM) is "Enhancements to the AIX Kernel for Support of Real-Time Applications", by Kathy Bohrer and John O'Quin, in _IBM RISC System/6000 Technology_, an IBM pub that collected several RS/6000 and AIX papers from IBM journals. Note that interrupt handlers, while preemptible (by kernel processes at higher priority), are not pageable and presumably not reentrant. As is generally the case in Unix, a poorly-written interrupt handler can hang or crash the system. Also, most of the synchronization primitives in the kernel are cooperative, at least in the AIX 3 design. (I don't know if AIX 4 enforces them.) There are plenty of potential pitfalls in this sort of implementation, but it's a start. > >AIX, like some other modern Unix implementations, doesn't have a > >separate pager process. All disk I/O (except I/O to special > >partitions that are reserved to special third-party applications, > >like DBMSes, and largely ignored by the OS) goes through the kernel's > >virtual memory manager. That means all swap and file I/O is > >memory-mapped, under the covers ("covers" being the read, write, etc. > >system calls) if necessary. > > OK, so I'm assuming that this VM manager is a bottleneck. And that > doesn't sound very good for performance (understand, that I come > from a group who was terribly fussy about good performance w.r.t > the user). It also means that general I/O is funneled through > the VM code. That just gives me the shivers--of course, it could > be that I'm biased to the hilt :-). Performance is fine, unless you run out of virtual memory; but thrashing is thrashing, and AIX does have a mechanism for recovering more or less gracefully from memory exhaustion (it starts killing off processes in some "least-favored" order, first with a catchable signal so the process can exit cleanly and then with an uncatchable one to enforce it, until it's below the danger mark). The VMM controls all the RAM and disk storage. It deals with pages, and it manages them using the hardware TLBs and inverse TLBs. When a process wants a page, chances are that it's already available in RAM, because RAM is effectively an MRU cache of pages. If it's not, the process gets preempted while the VMM schedules a page steal and page-in. If a process writes to a file, that page (or those pages) are marked for commitment (they'll be committed when someone issues a sync against that file or against the whole system; that happens periodically in the background if no user process does it explicitly). If another process tries to read the same part of that file, they'll get the copy from memory - the VMM knows that that file page is in RAM already. The VMM isn't generally a bottleneck because generally the most frequently accessed parts of the combined working sets of running processes are in RAM already. I have a number of Unix workstations I use daily (I develop and maintain a Micro Focus product on several Unix flavors), and my RS/6000s pause least frequently for disk I/O, be it file or swap. They don't have more RAM than my other machines; they just balance the scheduling better. Also, the kernel supports pinning pages in RAM, so the most frequently used kernel tables and whatnot never get paged out. > >The VMM, which is a kernel lightweight process, won't block forever > >on I/O problems because it has a watchdog timer. > > Good gads, it shouldn't! The requestor should probably have the > timeout for retries or whatever is appropriate for that particular > device medium (note that this should be at the device driver level). I should emphasize that much of what I'm saying here is speculative; the references I'm using don't go into a great deal of detail. Here's my understanding: The requestor makes its request and waits. It can be preempted (and in this case will be). At some point it gets back on the processor and sees a return code from the VMM: either success or failure. It can retry if it likes. It doesn't have to manage any timers. But again, that's just my impression of how it works, from reading various articles on it. I haven't looked at the source or written any kernel services (in AIX; I've done a little kernel hacking in BSD and Linux) myself. I could be completely wrong about timeouts being done in the VMM itself. > >> Assuming it is a temporary error (whatever that means...there are > >> myriads of categories), do you tell someone about the error? And, > >> if you do think someone should be notified, who? > > > >In Unix, that's a configuration issue, if the problem is non-fatal; > >it's logged via the syslog facility or something similar in most > >flavors, and you have various options for what "logging" involves. > > But what piece defines fatal? The kernel. Some piece of kernel code decides to call the panic service. It might be a device driver or some other sort of extension, or it might be part of the basic kernel shipped with the OS. > >If the kernel panics, that's a different story, but again details > >are implementation-specific. > > You just confused me. Which implementation? Site, user, OS, > hardware? How kernel panics are handled is generally a question of OS implementation (how your flavor of Unix implements panicking). > >Typically the kernel dumps core to > >the save device, if it can, and shuts down or reboots. Many modern > >machines have some facility for storing some error information in > >non-volatile memory, too. > > TOPS10 also had a "dump core and continue" stopcode category. That > was the basis for implementing the "suspend" option of Symmetric > Multi-Processing. The problem with Unix is that the panic may have been issued by third-party device driver code or kernel extensions. Who's to say what's been damaged? The typical Unix kernel does not provide a lot of protection against kernel code going haywire. Fault-tolerant Unix "systems" are generally clusters of semi-independent machines that monitor for one another and do takeovers when necessary. Good modern Unix implementations (like AIX) should handle panicking and automatically rebooting fairly well, in the sense of not sustaining filesystem damage (AIX uses a journaled filesystem), restarting the appropriate applications (ie., the ones configured for autostarting) and so on. But user processes will bounce and uncommitted data will be lost. > >> >Alternately, that disk I/O call could complete after passing the data > >> >off to the cache subsystem. The cache subsystem would then wait > >> >patiently for the disk to become available again. > >> > >> Thus you leave the user or program with the understanding that the > >> data has been written in iron. Now the next action by the same > >> user (or, more likely, another user) is to read back the newly > >> written data. How do you ensure data integrity? > > > >With the all-I/O-via-VMM architecture, it's not a problem. The OS > >knows which pages are in RAM and which are not. > > It is a problem if you wanted to do symmetric multi-processing. It > doesn't seem to allow I/O that is synchronous with the user requests > but uses an asynchonous process. (Or, I'm simply not understanding > what you're describing.) There's a mechanism for forcing synchronous, write-through I/O, for applications that need it. But even without it, user processes running on SMP systems sharing file data will see each other's updates, because each process is looking at the same pages - whether they're cached in RAM or not. > >But as you say that's not an ideal architecture for all applications, > >which is why you have DBMSes that write to raw partitions and don't > >go through the OS' VMM. > > And that sentence puts shivers of dread all up and down my spine. > This was an anathema in TOPS10 :-). I should make it clear that applications don't write directly to devices; they still go through system calls to the kernel device drivers and so on. They just have a faster path that doesn't use the filesystem code or the VMM. And the partitions they're writing to do NOT contain a user filesystem. They're only for application data. (Of course, that may still be anathema.) Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University However, we maintain that our mission is more than creating high-tech amusement--rather, we must endeavor to provide high-tech, high-touch entertainment with an emphasis on enkindling human warmth. -- "The Ultimate in Entertainment", from the president of video game producer Namco ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: jmfbahciv@aol.com's message of Thu, 10 Dec 98 13:02:30 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> Date: Thu, 10 Dec 1998 20:05:36 GMT Lines: 50 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!ix.netcom.com!netcom!alderson In article <74oi4r$nrp$4@ligarius.ultra.net> jmfbahciv@aol.com writes: >>AIX, like some other modern Unix implementations, doesn't have a separate >>pager process. All disk I/O (except I/O to special partitions that are >>reserved to special third-party applications, like DBMSes, and largely >>ignored by the OS) goes through the kernel's virtual memory manager. That >>means all swap and file I/O is memory-mapped, under the covers ("covers" >>being the read, write, etc. system calls) if necessary. >OK, so I'm assuming that this VM manager is a bottleneck. And that doesn't >sound very good for performance (understand, that I come from a group who was >terribly fussy about good performance w.r.t the user). It also means that >general I/O is funneled through the VM code. That just gives me the >shivers--of course, it could be that I'm biased to the hilt :-). Could be, Barb. ;-) After all, that's the I/O model used in Tops-20, and its predecessor TENEX, hardly "modern" in the sense the youngsters mean it. Oh, by the way, the I/O performance was good on DEC disks, and is still good on SCSI-2 disks from an old DEC customer. (How many people these days know that Hewlett-Packard used a Stanford-originated CAD package that ran on a DEC-20?) >>The VMM, which is a kernel lightweight process, won't block forever on I/O >>problems because it has a watchdog timer. >Good gads, it shouldn't! The requestor should probably have the timeout for >retries or whatever is appropriate for that particular device medium (note >that this should be at the device driver level). I think some confusion has set in: The VM code in this case *is* the requestor for services from an appropriate disk driver, and will time out eventually if the driver does not come back to it. >>With the all-I/O-via-VMM architecture, it's not a problem. The OS knows >>which pages are in RAM and which are not. >It is a problem if you wanted to do symmetric multi-processing. It doesn't >seem to allow I/O that is synchronous with the user requests but uses an >asynchonous process. (Or, I'm simply not understanding what you're >describing.) That may be why Tops-20 can never be made SMP; I've had it explained to be in the past, but since I was only ragging on someone about fiddles in the OS to start with, I didn't try to remember all the details of the explanation. :-) -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> <36708455.93118759@news.bright.net> From: Marco S Hyman Date: 10 Dec 1998 20:33:26 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 913350806 220 marc@204.94.187.130 Lines: 12 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!diablo.theplanet.net!dispose.news.demon.net!demon!woodstock.news.demon.net!demon!newsfeed.berkeley.edu!news2.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail deke.spamblock@generous.net writes: > version of cc. Can you name a Unix that does not contain cc? picking nits... for quite a while cc was an optional, extra-cost item that came with the 'developers' add-on. Interesting, in that a brain damaged (and hidden) version of the compiler and linker were still provided to configure and build a custom kernel. See, for example, Interactive UNIX Sys V R3 and R3.2. Yes, it was sold as UNIX, not a UNIX-like operating system :-) // marc ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 21:14:36 -0600 Organization: MindSpring Enterprises Lines: 32 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> NNTP-Posting-Host: d1.56.d1.38 X-Server-Date: 11 Dec 1998 03:14:39 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <367551c8.59800829@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Can they, in any way, be considered as files themselves? >In that case you could say they're hacked onto a flat file >system, just like in every OS I know. Yes, but I don't agree that that makes the arrangement a flat file system. Read on. >Or is all information, including that needed to determine what >directory an entry belongs to, stored in one single giant bag? -That- would be a flat file system, such as MFS (Macintosh File System, the original Macintosh file system). One name space, with "folders" kinda kludged on top, but no file or folder names could be the same. On VMS, a directory is a file. You can open it, read it, rename it (though that's not a great idea), and delete it, all similarly to a normal file. If a directory is empty, you can delete it w/o any consideration to the fact that it could "contain" files. The main difference here between UNIX and VMS, as I see it, is that files on VMS are strongly typed. File attributes beyond size and dates are held by the file header. While UNIX has some file attributes (I'm not real solid on this), there appear to be relatively few of them, and they are very general. VMS gets very specific, tracking information such as carriage control, record length, modification count (this is not the version number), and other things. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 21:18:36 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> NNTP-Posting-Host: d1.56.d1.38 X-Server-Date: 11 Dec 1998 03:18:38 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74nqnu$1si$1@news.latrobe.edu.au>, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >You can always do this with VMS too. See SET FILE/ENTER. To be fair, this is not a symbolic link. This is a hard link. As on UNIX, it can't cross disk ("partition" on UNIX) bounderies. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 21:31:33 -0600 Organization: MindSpring Enterprises Lines: 59 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> <74p71s$86r@news3.newsguy.com> NNTP-Posting-Host: d1.56.d1.38 X-Server-Date: 11 Dec 1998 03:31:37 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74p71s$86r@news3.newsguy.com>, mww@microfocus.com (Michael Wojcik) wrote: >I don't think my Tom Lehrer CD has "New Math". It has "Lobachevsky", >though. "New Math" has Lehrer (sp?), a math professor, doing an arithmetic problem from a text book using the audience as "students". After he does the problem, he does it again after announcing that the book wanted it done in "base 8". Which is, as someone else quoted, "just like base 10... if you're missing two fingers!" >Yes, I was a bit out of line there. Maybe, but let's just say you were provoked. >Unix is really targetted toward >the kinds of programmers the _New Hacker's Dictionary_ calls >"toolsmiths" - people who like to pick up odd things here and there >and put them together. It's a game. If you're not in the mood to >play it, it's bound to be annoying. GOOD summation. I'll have to remember that. I've been a "toolsmith", but I haven't been able to indulge that desire much since I left school. I -did- recently design a tool... but it's for a single user and that user continues to not use it. It's not been QA qualified yet. >It's tough to specify that across hardware lines. Support for >portability could be better, though. The system headers for >things like symbol table entries tend to be a bit opaque. (I added >debugging symbol generation to a compiler running under SunOS 4 >some years back, and spent a fair bit of time compiling little C >fragments to assembly to see what a "real" compiler did.) Define "real compiler", huh? Just because you worked on it doesn't mean it's less "real" than a compiler built by the compiler writers in the back room, who are larger than life because you never meet them. Anyway, an important point we both missed in my description of the LDPCTX/SVPCTX instructions is that the line between hardware function and software function is kinda blurry. Granted, if everyone was to emulate the VAX calling standard, it would make programs on architectures drastically different from VAX less efficient, but it -would- be possible. >I'd rate >the AS/400 slightly higher because of its additional levels of >abstraction and object orientation (even CISCier than VAX), but it >came much later and benefitted from years of advances in hardware >engineering, which left a lot of headroom; and even so the first >400s were abysmally slow at the low end. Haven't gotten my hands on one of these, and I probably never will. So be it. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 21:34:45 -0600 Organization: MindSpring Enterprises Lines: 30 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> <74o7jt$6d6$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: d1.56.d1.38 X-Server-Date: 11 Dec 1998 03:34:47 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74o7jt$6d6$1@news.wlg.netlink.net.nz>, don@news.daedalus.co.nz (Don Stokes) wrote: >In article , >Howard S Shubs wrote: >>I have to yield this one. I seem to remember that VAX has a tracepoint >>bit, and watchpoints were implemented in some way I wouldn't have thought >>of, but that they were both effectively hardware. > >It's done with page protections -- the page containing the watchpoint is >marked read-only, and if the code attempts to modify it, [etc.] > >I might be a little off on fine detail, but you get the idea. As I said, in some way I wouldn't have thought of. >That mechanism can be done on pretty much anything with paged memory >management and OS services permitting the debugger to futz with the >page protections. Which is my complaint. >(Note that under VMS, the guts of the debugger run >in the context of the process running the code being debugged.) So what's the separate debugger process? Just a front-end? -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 21:36:21 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <74no1h$19v$1@news.latrobe.edu.au> <74ogik$nrp$1@ligarius.ultra.net> NNTP-Posting-Host: d1.56.d1.38 X-Server-Date: 11 Dec 1998 03:36:24 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74ogik$nrp$1@ligarius.ultra.net>, jmfbahciv@aol.com wrote: >You may keep it. I didn't touch VMS (unless forced to) >because of being offended. NT is far, far more advanced >in that category. So -that- is what "new technology" NT is! It's More Annoying! -- Howard S Shubs The Denim Adept ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk> From: Marco S Hyman Date: 10 Dec 1998 21:43:04 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 Lines: 30 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 913354992 220 marc@204.94.187.130 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news2.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail hshubs@mindspring.com (Howard S Shubs) writes: > No, neither are they available on UNIX. Far as I can tell, the only paper > manuals available for UNIX are compilations of MAN pages. There are no In general, or for a specific program? In general non-man documentation is founnd in /usr/share/doc (at least for *bsd unix). Admittedly, there are more chapters referenced than there exist actual files. And I'd guess that much of the doc that is there is probably out of date. But anyway, try this: cd # your home dir mkdir doc cd doc lndir /usr/share/doc # so you don't polute /usr for for d in smm psd usd; do (cd $d && make); done and the result is a handfull of postscript files suitable for printing. On my system, the result is: ./psd/05.sysman/paper.ps ./smm/01.setup/paper.ps ./usd/04.csh/paper.ps ./psd/12.make/paper.ps ./smm/03.fsck_ffs/paper.ps ./usd/18.msdiffs/paper.ps ./psd/18.gprof/paper.ps ./smm/04.quotas/paper.ps ./usd/19.memacros/paper.ps ./psd/19.curses/paper.ps ./smm/05.fastfs/paper.ps ./usd/20.meref/paper.ps ./psd/20.ipctut/paper.ps ./smm/06.nfs/paper.ps ./psd/21.ipc/paper.ps ./smm/07.lpd/paper.ps ./smm/18.net/paper.ps // marc ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> <74p71s$86r@news3.newsguy.com> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 43 Message-ID: Date: Thu, 10 Dec 1998 22:50:32 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913330232 205.166.146.8 (Thu, 10 Dec 1998 16:50:32 CDT) NNTP-Posting-Date: Thu, 10 Dec 1998 16:50:32 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <74p71s$86r@news3.newsguy.com>, Michael Wojcik wrote: >I don't think my Tom Lehrer CD has "New Math". It has "Lobachevsky", >though. You probably have the "older" one. I am aware of three "old" albums and one "new" album. The "old" ones (which I listened to endlessly as a child) were * Songs by Tom Lehrer (keyboard with Tom Lehrer sitting at it, he has a devil's tail in the picture) * The Old Dope Peddler * The Irish Ballad * The Hunting Song * Generically disturbing material, no patter * An Evening Wasted With Tom Lehrer (picture of audience, and piano with spotlight) * New Math * Clementine * Funny songs with introductions and a laugh track * That Was The Year That Was (headlines) * Whatever Became of Hubert * Who's Next * News-related songs The more recent one is the "More Songs by Tom Lehrer", and contains most of the tracks from the Songs By and An Evening Wasted With, as well as a couple of versions with orchestral accompaniment. It omits most of the patter and discussion, but does have almost all of the songs. I probably have many of the details wrong; I have all three of these (I have the "More Songs" version), but I can't imagine where I could have put them. I've never been one for finding things. New Math is one of my favorites, and is a lot funnier now than it was before I used octal regularly. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 10 Dec 1998 23:43:13 GMT Organization: La Trobe University Lines: 19 Message-ID: <74pmah$bsc$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74nr8t$1si$2@news.latrobe.edu.au> <74p06d$oso$1@agate.berkeley.edu> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 913333393 12172 131.172.12.11 (10 Dec 1998 23:43:13 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 10 Dec 1998 23:43:13 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!enews.sgi.com!harbinger.cc.monash.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Eric J. Korpela (korpela@islay.ssl.berkeley.edu) wrote: : What exactly do clusters give you that NFS+(YP || NIS) doesn't give you? Reliability. : An artificially short maximum path between computers? (Maybe this has : changed since I used VMS, but machines in a cluster needed to be basically : on the same subnet.) At one point I worked on the (physically) largest : VAX cluster in the world, which had a 25km optical fiber link between : two nodes. That was a long, long time ago, though.) I believe the limit is in excess of 300Km which is large enough for most applications. -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 23:49:16 GMT Organization: . Lines: 21 Message-ID: <367551c8.59800829@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> NNTP-Posting-Host: pool02b-194-7-150-33.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news.gsl.net!gip.net!newsfeed.tli.de!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach hshubs@mindspring.com (Howard S Shubs) on Wed, 09 Dec 1998 22:51:59 -0600 to alt.folklore.computers: > In article <74mabs$rrl$1@agate.berkeley.edu>, > korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: > > >Strange. Are you telling me that VMS directories weren't hacked onto a > >flat file system? > > Can't say. They don't -look- like it. Can they, in any way, be considered as files themselves? In that case you could say they're hacked onto a flat file system, just like in every OS I know. Or is all information, including that needed to determine what directory an entry belongs to, stored in one single giant bag? -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 10 Dec 1998 23:49:16 GMT Organization: . Lines: 34 Message-ID: <36775424.60404627@news.uunet.be> References: <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> NNTP-Posting-Host: pool02b-194-7-150-33.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news.gsl.net!gip.net!newsfeed.tli.de!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach alderson@netcom.netcom.com (Richard M. Alderson III) on Wed, 9 Dec 1998 19:37:17 GMT to alt.folklore.computers: > On HP-UX systems, there is a system management tool called SAM, the binaries > for which live in a directory on /usr. Now, on any decent system, /usr is > going to live on a different filesystem than /, so in single-user mode it may > not be available--like when moving /usr to a larger filesystem. > > So I, at least, put a copy of /usr/sam (and a couple of other tools) in the > /usr directory which is the mountpoint for the /usr filesystem. Made my life > a lot simpler, and I hardly think it was stupid--it became SOP for a large > (then $750,000,000) networking company I worked for at the time... Somewhere - I'm not sure anymore, but I think it was in a Linux system administrator's manual (read "file sa."), I saw a recommendation to do something like this if you want to put /etc on another than the boot volume: put everything that's necessary for booting in a minimal /etc, and later mount the 'real' /etc over it. I've been asking myself serious questions about this approach, but I think it can be made relatively safe if you () on booting, copy the small /etc into the big one right before mounting it () on shutdown, copy those same files (only what you need) back after unmounting it. That is, if you don't need files that are different in the two /etc's to make it work (haven't thought that far - never needed to). -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: deke.spamblock@generous.net Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 02:51:00 GMT Organization: GenerousCity is a virtue - find romance at http://generous.net Lines: 41 Message-ID: <36708455.93118759@news.bright.net> References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: paul-cas1-cs-38.dial.bright.net X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!204.59.152.222!news-peer.gip.net!news.gsl.net!gip.net!news.new-york.net!news-feed.fnsi.net!nntp.msen.com!news.altair.com!NewsNG.Chicago.Qual.Net!news.bright.net!not-for-mail On 10 Dec 1998 00:57:30 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >I'd say MJB is a little more authoritative than the SCO documentation, >even in the Unix community. Well, make up your mind. You were the one that was making such a big deal about branding. SCO is a licensed publisher of branded Unix. >>But what does that have to do with anything? Solaris is not "straight >>unix" - it has a number of features that other unixes do not have. >>That's why a program will say "requires Solaris to run" instead of >>"requires any Unix operating system." >You're totally missing the point. I'm talking about basic features >such as #!/bin/sh script. I have never seen a #!/bin/sh script that >said "runs on Unix systems with /bin/sh installed". Rather they say >"runs on Unix system" because /bin/sh is part of Unix. I never said the Bourne shell was not part of Unix. The question did not talk of *all* Unix. It talked about *straight* Unix. There are *lots* of shell scripts that refuse to work on the bourne shell. Just as there are lots of c source files for unix that refuse to compile on Sun's version of cc. Can you name a Unix that does not contain cc? Shells are application programs, and users can AND DO pick the shell they want, just like they choose whether they want to edit files with emacs, joe, or vi, or they choose between using more or less. The fact that a car comes with a "space-saver spare" doesn't mean that it stops being a car if you use a *real* tire for a spare - or if you eliminate the spare entirely. deke ------------------------ Let love find you! http://generous.net A list for flirting GenerousSingles-subscribe@onelist.com Over The Hill Gang GenerousSinglesOTHG-subscribe@onelist.com Personal Ads list GenerousProfiles-subscribe@onelist.com ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> <36708455.93118759@news.bright.net> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 25 Message-ID: Date: Fri, 11 Dec 1998 05:14:52 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913353292 205.166.146.8 (Thu, 10 Dec 1998 23:14:52 CDT) NNTP-Posting-Date: Thu, 10 Dec 1998 23:14:52 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <36708455.93118759@news.bright.net>, wrote: >There are *lots* of shell scripts that refuse to work on the bourne shell. Just >as there are lots of c source files for unix that refuse to compile on Sun's >version of cc. Can you name a Unix that does not contain cc? Well, Solaris, supposedly. Sun's "cc" is not a C compiler. I am unaware of any shell scripts that don't run under sh, although I've seen lots of ksh or csh scripts that don't run under sh. :) >they choose between using more or less. The fact that a car comes with a >"space-saver spare" doesn't mean that it stops being a car if you use a *real* >tire for a spare - or if you eliminate the spare entirely. True. However, 'sh' is around the level of 'gas tank'. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### bject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 08:07:32 -0600 Organization: MindSpring Enterprises Lines: 19 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> <74qlg2$g5m$1@news.latrobe.edu.au> NNTP-Posting-Host: d1.56.d0.5d X-Server-Date: 11 Dec 1998 14:07:33 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74qlg2$g5m$1@news.latrobe.edu.au>, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >Howard S Shubs (hshubs@mindspring.com) wrote: >: In article <74nqnu$1si$1@news.latrobe.edu.au>, >: cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: > >: >You can always do this with VMS too. See SET FILE/ENTER. > >: To be fair, this is not a symbolic link. This is a hard link. As on >: UNIX, it can't cross disk ("partition" on UNIX) bounderies. > >Agreed, and I don't recommend that anyone actually use SET FILE/ENTER >to do this sort of thing, but it's a bit of VMS file funtionality >that lots of people don't know about... And that's good. How many people know about the CHECKSUM command, either? -- Howard S Shubs The Denim Adept ###### From: cchd@lucifer.its.latrobe.edu.au (Huw Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 08:35:14 GMT Organization: La Trobe University Lines: 17 Message-ID: <74qlg2$g5m$1@news.latrobe.edu.au> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> NNTP-Posting-Host: lucifer.its.latrobe.edu.au X-Trace: news.latrobe.edu.au 913365314 16566 131.172.12.11 (11 Dec 1998 08:35:14 GMT) X-Complaints-To: usenet@news.latrobe.edu.au NNTP-Posting-Date: 11 Dec 1998 08:35:14 GMT X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!news.mel.connect.com.au!news.syd.connect.com.au!nsw.nntp.telstra.net!news1.mpx.com.au!news1.optus.net.au!optus!vrn.edu.au!news.swin.edu.au!news.latrobe.edu.au!lucifer.its.latrobe.edu.au!cchd Howard S Shubs (hshubs@mindspring.com) wrote: : In article <74nqnu$1si$1@news.latrobe.edu.au>, : cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: : >You can always do this with VMS too. See SET FILE/ENTER. : To be fair, this is not a symbolic link. This is a hard link. As on : UNIX, it can't cross disk ("partition" on UNIX) bounderies. Agreed, and I don't recommend that anyone actually use SET FILE/ENTER to do this sort of thing, but it's a bit of VMS file funtionality that lots of people don't know about... -- Huw Davies | e-mail: Huw.Davies@latrobe.edu.au Information Technology Services | Phone: +61 3 9479 1550 Fax: +61 3 9479 1999 La Trobe University | "If God had wanted soccer to be played Melbourne Australia 3083 | in the air, the sky would be green" ###### From: huge@nospam.demon.co.uk (Hugh Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 09:42:21 GMT Organization: Piglet's Pickles and Preserves Message-ID: <74qpdt$57u@axalotl.demon.co.uk> References: <1456.648T1620T6105137@sky.bus.com> Reply-To: huge@nospam.demon.co.uk NNTP-Posting-Host: axalotl.demon.co.uk X-NNTP-Posting-Host: axalotl.demon.co.uk:158.152.24.143 X-Trace: news.demon.co.uk 913415880 nnrp-07:6849 NO-IDENT axalotl.demon.co.uk:158.152.24.143 X-Complaints-To: abuse@demon.net x-no-archive: yes Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!axalotl.demon.co.uk!axalotl!usenet In article hshubs@mindspring.com (Howard S Shubs) writes: >Granted. Still, when I had to get an economist and an accountant used >to UNIX (which isn't yet a completed project), I had -lots- of fun >explaining certain things, like the chmod protection mask. So don't do it that way. What's wrong with chmod u+x,go-rwx filename ??? -- "The road to Paradise is through Intercourse." The uk.transport FAQ; http://www.axalotl.demon.co.uk/transport/FAQ.html [Substitute "axalotl" for "nospam" to email me] ###### From: jkatz@ultrix6.cs.csubak.edu (who?) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 09:46:01 GMT Organization: California State University, Bakersfield Lines: 40 Message-ID: <74qpkp$rm5$1@hades.csu.net> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> NNTP-Posting-Host: ultrix6.cs.csubak.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!usc!newshub.csu.net!ultrix6.cs.csubak.edu!jkatz Robert Billing (unclebob@tnglwood.demon.co.uk) wrote: : Howard S Shubs wrote: : > Logical names are, in part, a way for system managers to alias things in : > order to organize the system better. As a UNIX-head, you're familiar with : > the concept of the "~" directory. Which is better, using that notation or : > using the full path? : The other virtue of logical names is that they can be different for : different users. It is perfectly possible to set up the system in such a : way that the system login file creates logical names for each session : which depend on the user ID, time, phase of the moon, number of : [mongooses|mongeese] in stock... : The virtue of symlinks is that they are the same for all users, : regardless of user ID... : The two systems are complimentary, and implementing logical names in : un*x, as well as symlinks, would be a Good Thing. There seems to be a lot of confusion concerning *VMS logicals. To compare them to symlinks is, at least in my mind, erroneous. They are more akin to global variables, if you were to consider shell variables and DCL symbols to be local variables. A bit like perl's hash arrays, they allow programs to look up a value given a name. As it happens, they are frequently used to refer to parts of the file system, but there is no reason why they can't contain a network address, or a terminal device. Like shell vars, it's up to the program to decide what they represent. There is something akin to links in the ODS-2 file system, but they are a bit on the dangerous side, as it is possible to remove all references to a file, without marking the space as available. for UNIX links, the rm doesnt free the space untill the last link is gone. VMS has a seperate command to remove a link. I have no idea what happens if you DELETE a file while other links remain. just had to get that out jeremy ###### From: huge@nospam.demon.co.uk (Hugh Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 09:53:58 GMT Organization: Piglet's Pickles and Preserves Message-ID: <74qq3m$57u@axalotl.demon.co.uk> References: Reply-To: huge@nospam.demon.co.uk NNTP-Posting-Host: axalotl.demon.co.uk X-NNTP-Posting-Host: axalotl.demon.co.uk:158.152.24.143 X-Trace: news.demon.co.uk 913415883 nnrp-07:6849 NO-IDENT axalotl.demon.co.uk:158.152.24.143 X-Complaints-To: abuse@demon.net x-no-archive: yes Lines: 23 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!axalotl.demon.co.uk!axalotl!usenet In article , seebs@plethora.net (Peter Seebach) writes: >In article <74p71s$86r@news3.newsguy.com>, >Michael Wojcik wrote: >>I don't think my Tom Lehrer CD has "New Math". It has "Lobachevsky", >>though. > >You probably have the "older" one. I am aware of three "old" albums and >one "new" album. The "old" ones (which I listened to endlessly as a child) >were > * Songs by Tom Lehrer (keyboard with Tom Lehrer sitting at it, he has > * An Evening Wasted With Tom Lehrer (picture of audience, and piano > * That Was The Year That Was (headlines) Hmmm. I have "Another Evening Wasted with Tom Lehrer", as well. -- "The road to Paradise is through Intercourse." The uk.transport FAQ; http://www.axalotl.demon.co.uk/transport/FAQ.html [Substitute "axalotl" for "nospam" to email me] ###### From: jkatz@ultrix6.cs.csubak.edu (who?) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 09:54:42 GMT Organization: California State University, Bakersfield Lines: 26 Message-ID: <74qq52$rm5$2@hades.csu.net> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: ultrix6.cs.csubak.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!usc!newshub.csu.net!ultrix6.cs.csubak.edu!jkatz TERENCE MURPHY (tsmurphy@cs.uiuc.edu) wrote: : In article <74kofa$8tf@news1.newsguy.com>, : Michael Wojcik wrote: : >The shell, Bourne or otherwise, isn't part of "straight Unix", if : >by that we mean the core OS. : The shell is part of POSIX.2 (the second and third letters stand for : "operating system") and also part of the OpenGroup Unix specification. : If you delete /bin/sh what you have left will not be Unix (since it : will not pass the POSIX or OpenGroup specifications). Careful... poking about DU 4.0, when, I stumbled upon this: myhost:~$ diff /bin/sh /usr/bin/posix/sh Binary files /bin/sh and /usr/bin/posix/sh differ Now, Im assuming the second is the posix compliant shell. DU is indeed UNIX branded. Now as to the issue of whether Unix (you used 3 lower case letters) is a brand granted by (err, who owns it now?), or the description of an operating environment, I wont argue about that (not now anyway). cheers, jeremy ###### From: jkatz@ultrix6.cs.csubak.edu (who?) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 09:58:15 GMT Organization: California State University, Bakersfield Lines: 20 Message-ID: <74qqbn$rm5$3@hades.csu.net> References: <73ela0$51h@hermes.acs.unt.edu> <74kasb$47f$1@mathserv.mps.ohio-state.edu> <366df034.14353548@news.xmission.com> <74m80v$4od$1@vixen.cso.uiuc.edu> <74ma75$eno$1@pegasus.csx.cam.ac.uk> NNTP-Posting-Host: ultrix6.cs.csubak.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!howland.erols.net!usc!newshub.csu.net!ultrix6.cs.csubak.edu!jkatz Ben Harris (bjh21@cus.cam.ac.uk) wrote: : In article <74m80v$4od$1@vixen.cso.uiuc.edu>, : TERENCE MURPHY wrote: : >Yes, but there is free VMS for VAX (peecee's are boring as hell : >anyways, why would you want to use one?) Just go to www.ebay.com, pick : >yourself up a VAXstation (for $200), then go to www.montagar.com and : >pick up a VMS license & media (for $30), : ... assuming you're a member of one of the relevant DECUS chapters. As far : as I can see, that'd entail my moving to the Netherlands at least, which is : probably more effort than I'd like to go to for a VMS licence. TKFAD has offered now for at least a few months, although rather quietly, a hobbyist license outside of the one through DECUS. I dont know if there is any restriction in the department of nations. It is limited to workstation clas use on VAXen, but allows the latest release. email me if you'de like a URL for reference. Address should be in the headers. jeremy ###### From: jkatz@ultrix6.cs.csubak.edu (who?) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 10:02:00 GMT Organization: California State University, Bakersfield Lines: 34 Message-ID: <74qqio$rm5$4@hades.csu.net> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <366E74AB.6EC53A58@trailing-edge.com> NNTP-Posting-Host: ultrix6.cs.csubak.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!usc!newshub.csu.net!ultrix6.cs.csubak.edu!jkatz Tim Shoppa (shoppa@trailing-edge.com) wrote: : Eric J. Korpela wrote: : > : > In article , : > Howard S Shubs wrote: : > >In article <74jpov$4iu$1@agate.berkeley.edu>, : > >korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: : > > : > >>The VMS file system has always felt to me like it is something cobbled onto : > >>a flat file system. : > > : > >Strange. Probably due to a lack of experience with VMS vs some experience : > >with RSX-11. : > : > Strange. Are you telling me that VMS directories weren't hacked onto a : > flat file system? : Only in the same sense that Unix directories were hacked onto : a flat file system. With either, there's always a disk : that's a "flat" collection of blocks, accessed by block number : (whether you choose to call them LBN's, VBN's, or inodes.) : (Yes, I know there are computers where mass disk storage is done : associatively, and not by block number. But I don't know of any : Unix or VMS systems that do so.) Actually, I was reading a description that D. Ritchie posted to a site at Bell Labs concerning early UNIX that described the file system as something about as primitive. The solution used was the now disallowed used of hard links to a single directory which allowed sub directories. If you spend too much time thinking about ancestry, you're sure to come up with something worth not boasting too loudly about. jeremy ###### Date: 11 Dec 98 10:34:19 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> <36708455.93118759@news.bright.net> Message-ID: <1311.649T954T6343891@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 27 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 11 Dec 1998 13:33:56 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!128.230.129.106!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.120 In article marc@snafu.org (Marco S Hyman) writes: >picking nits... for quite a while cc was an optional, extra-cost >item that came with the 'developers' add-on. Interesting, in that >a brain damaged (and hidden) version of the compiler and linker >were still provided to configure and build a custom kernel. See, >for example, Interactive UNIX Sys V R3 and R3.2. Yes, it was sold >as UNIX, not a UNIX-like operating system :-) This sounds like the assembler for Sperry->Unisys OS/3. Partway through OS/3's lifetime, the assembler was "unbundled" (i.e. leased for a ridiculous charge when it once was free), but you got a copy whether you paid or not because you needed it to do sysgens. You just had to promise not to use it for anything else. This, along with removing the documentation for physical IOCS from the appropriate manuals, reinforced my belief that Sperry was doing its best to kill assembly language programming. (I made sure to hold on to an old copy of the manuals so I could continue to do my various illicit hacks such as device-independent programs.) Once I got my own assembler written, I got a lot of joy out of giving a copy to anyone who wanted it. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 98 11:25:27 GMT Organization: UltraNet Communications, Inc. Lines: 17 Message-ID: <74r0r2$o32$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <747c79$jo7$1@news.latrobe.edu.au> <36677771.23568198@news.vip.net> <748j4a$lfi$2@strato.ultra.net> <74no1h$19v$1@news.latrobe.edu.au> <74ogik$nrp$1@ligarius.ultra.net> NNTP-Posting-Host: d14.dial-16.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 11 Dec 1998 11:48:50 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!feed1.news.rcn.net!rcn!news.ultranet.com!d14 In article , hshubs@mindspring.com (Howard S Shubs) wrote: >In article <74ogik$nrp$1@ligarius.ultra.net>, jmfbahciv@aol.com wrote: > >>You may keep it. I didn't touch VMS (unless forced to) >>because of being offended. NT is far, far more advanced >>in that category. > >So -that- is what "new technology" NT is! It's More Annoying! I didn't use the word annoying. One can work with annoying; one can't work with offending. /BAH Subtract a hundred and four for e-mail. ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 13:29:26 GMT Organization: Daedalus Consulting Lines: 28 Message-ID: <74r6nm$gq1$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74o7jt$6d6$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article , Howard S Shubs wrote: >As I said, in some way I wouldn't have thought of. I put it in the "clever, but kinda obvious when you think about it" category. >So what's the separate debugger process? Just a front-end? Yep. Mainly, having the front-end of the debugger in its own process means that part of it can run without accidentally stomping on bits of the process being debugged, and vice-versa. This is particularly useful for programs using the SMG (screen management) library, since the debugger also uses SMG. The debugger used to live entirely within the process being debugged. I think it changed around V5. (V4.4 was when the debugger started using SMG -- before that it did its own screen management. There were a bunch of other enhancements then too, but I'm pretty sure the two-process arrangement came later.) Stomping on the debugger used to be a real hassle -- I've even been known to put watchpoints on what I knew to be debugger data structures, because in the previous run that's where the debugger crashed after a (inevitably C) program followed a wayward pointer. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 13:45:28 GMT Organization: Daedalus Consulting Lines: 27 Message-ID: <74r7lo$l7g$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Huw Davies wrote: [symbolic links] >You can always do this with VMS too. See SET FILE/ENTER. SET FILE/ENTER is *not* a symbolic link. It's closest unix analogue is a hard link, and like unix hard links, it can't span filesystems. But it differs from unix hard links in that VMS does not keep a reference count of links to files as unix does, so if you delete a file that has a hard link to it from somewhere else, the link points to a non-existent file (ie you'll get a "no such file" error if you attempt to open it), whereas under unix, a file delete simply deletes the link, decrements the link reference count, and only if the refcount goes to zeros does it actually delete the file. Symlinks can point to files that don't exist, but if you delete a symlink, the real file sticks around (assuming it existed in the first place). Later versions (post V5) play some funny games defining a "primary" and "secondary" links, defining the "primary" link as the one where the file header file name and backlink match the directory access for that particular link, and "secondary" links as those where it doesn't. I'm not entirely sure why the bothered. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: bjh21@cus.cam.ac.uk (Ben Harris) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 13:48:03 GMT Organization: University of Cambridge, England Lines: 17 Message-ID: <74r7qj$rbj$1@pegasus.csx.cam.ac.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> NNTP-Posting-Host: taurus.cus.cam.ac.uk Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!news-lond.gip.net!news.gsl.net!gip.net!insnet.net!uunet!in3.uu.net!news1-gui.server.ntli.net!news-feed.ntli.net!news5-gui.server.ntli.net!news-feed.ntli.net!server4.netnews.ja.net!server2.netnews.ja.net!pegasus.csx.cam.ac.uk!bjh21 In article <1djvj1d.8wyafb10bgojyN@n244-87.berlin.snafu.de>, Juergen Nickelsen wrote: >Some vendors may have added types, of which I knwo only Solaris's >"doors", yet another interprocess communication endpoint. 4.4BSD has "whiteouts", but they're not really files (having only a directory entry and no inode. >Several file attributes are slightly simpler: no Delete permission -- >who has write permission for a directory may delete files in it; no >separate access permissions for SYSTEM; no creation date. Erm, st_ctime? -- Ben Harris Computer Officer, Corpus Christi College, Cambridge. ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 14:38:51 +0100 Organization: [Posted via] Interactive Networx Lines: 43 Message-ID: <1djvj1d.8wyafb10bgojyN@n244-87.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> NNTP-Posting-Host: n244-87.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news-hh.maz.net!unlisys!news.snafu.de!jnickelsen Howard S Shubs wrote: > The main difference here between UNIX and VMS, as I see it, is that files > on VMS are strongly typed. File attributes beyond size and dates are held > by the file header. While UNIX has some file attributes (I'm not real > solid on this), there appear to be relatively few of them, and they are > very general. VMS gets very specific, tracking information such as > carriage control, record length, modification count (this is not the > version number), and other things. Unix files are also strongly typed, but there are indeed not many types and only one type for "regular" files. Other common types are: - directory - symbolic link - character device (such as a serial port) - block device (such as a disk) - named pipe - socket (Named pipes and sockets are probably best compared to VMS's "mailboxes", but since I don't know any specifics of the latter, I won't go into details. :-) Some vendors may have added types, of which I knwo only Solaris's "doors", yet another interprocess communication endpoint. For regular files there is only one type because Unix doesn't care at all about the structure of a file; this is left to the application. From Unix's view all files are an unstructured stream of bytes. Several file attributes are slightly simpler: no Delete permission -- who has write permission for a directory may delete files in it; no separate access permissions for SYSTEM; no creation date. No ACLs in "vanilla" Unix. Several vendors have added ACLs, but these are (I think) mutually incompatible and not well integrated into the system. They are also, as I have heard, not as flexible as VMS's (but I may be completely wrong here since I never used them). The Unix file system *is* a simpler world, yes. -- Juergen Nickelsen ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 18:23:34 GMT Organization: Laser-Scan Ltd. Lines: 14 Message-ID: <74rnv6$62s@relay.lsl.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> <74r7lo$l7g$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!diablo.theplanet.net!join.news.pipex.net!pipex!news.lsl.co.uk!benh Don Stokes (don@news.daedalus.co.nz) wrote: : Later versions (post V5) play some funny games defining a "primary" and : "secondary" links, defining the "primary" link as the one where the : file header file name and backlink match the directory access for that : particular link, and "secondary" links as those where it doesn't. I'm : not entirely sure why the bothered. It would allow a backup program, for example, to weed out extra links to a file and only store one copy. (Of course, there are several other ways to do this.) -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: lucvdv@null.net's message of Thu, 10 Dec 1998 23:49:16 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> Date: Fri, 11 Dec 1998 20:03:04 GMT Lines: 21 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!feed1.news.rcn.net!rcn!newsfeed.enteract.com!ix.netcom.com!netcom!alderson In article <36775424.60404627@news.uunet.be> lucvdv@null.net (Luc Van der Veken) writes: >I've been asking myself serious questions about this approach, >but I think it can be made relatively safe if you >() on booting, copy the small /etc into the big one right before >mounting it >() on shutdown, copy those same files (only what you need) back >after unmounting it. Unless Linux is very different from the Unices I've worked with (SunOS, HP-UX, Solaris, Ultrix, OSF/1, Digital Unix 4.0, Pyramid), you can't copy files to an unmounted filesystem, so this can't work. Just leave the files in both the mountpoint directory and the mounted filesys- tem--rarely will you need to update the "little" copy because you only use it in limited circumstances. -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 20:14:45 GMT Organization: . Lines: 18 Message-ID: <36726db1.1210039@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> <36708455.93118759@news.bright.net> NNTP-Posting-Host: pool02b-194-7-80-228.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach seebs@plethora.net (Peter Seebach) on Fri, 11 Dec 1998 05:14:52 GMT to alt.folklore.computers: > In article <36708455.93118759@news.bright.net>, > wrote: > >they choose between using more or less. The fact that a car comes with a > >"space-saver spare" doesn't mean that it stops being a car if you use a *real* > >tire for a spare - or if you eliminate the spare entirely. > > True. However, 'sh' is around the level of 'gas tank'. You mean it's what you use when the gas tank is empty? [sorry - couldn't resist] -- My undergraduate experience convinced me that I was not smart enough to be a physicist, and that computers were quite neat. Dennis M. Ritchie ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 20:14:45 GMT Organization: . Lines: 14 Message-ID: <36736f13.1564439@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <74knnt$8c1@news1.newsguy.com> <74o7jt$6d6$1@news.wlg.netlink.net.nz> <366FA28F.40D8C49E@tnglwood.demon.co.uk> NNTP-Posting-Host: pool02b-194-7-80-228.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach Peter Ludemann on 10 Dec 1998 10:50:46 -0800 to alt.folklore.computers: > And, BTW, for all of you who lament Unix's expansion of file names in > the shell (hmmm, why hasn't anybody mentioned DataGeneral's OS which > also had the program expand filenames, just like VMS?) And DOS? All you had to do was link in a different startup routine to have them expanded before you arrived in main(). -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 20:14:46 GMT Organization: . Lines: 35 Message-ID: <3674744b.2900901@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> NNTP-Posting-Host: pool02b-194-7-80-228.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach hshubs@mindspring.com (Howard S Shubs) on Thu, 10 Dec 1998 21:14:36 -0600 to alt.folklore.computers: > In article <367551c8.59800829@news.uunet.be>, lucvdv@null.net (Luc Van der > Veken) wrote: > > >Can they, in any way, be considered as files themselves? > >In that case you could say they're hacked onto a flat file > >system, just like in every OS I know. > > Yes, but I don't agree that that makes the arrangement a flat file > system. Read on. [I did, and I understand what you mean] It's just how you look at it. When a directory is just another file, you could say "it may have been flat at first, and they hacked the file system so that it can go for additional dir entries in a file, without changing much to the way an entry looks". When all information is in a single directory, it's more like "it would almost certainly require a complete redesign of the dir entries to make this out of a flat file system, so they probably made it like this right from the start". I think I've got a pretty good idea about which of the two paths would require the largest amount of new (or changed) code, and which one would be the easiest to keep backward compatible with older disks. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 20:14:47 GMT Organization: . Lines: 21 Message-ID: <367579ce.4311729@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> <74qlg2$g5m$1@news.latrobe.edu.au> NNTP-Posting-Host: pool02b-194-7-80-228.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach hshubs@mindspring.com (Howard S Shubs) on Fri, 11 Dec 1998 08:07:32 -0600 to alt.folklore.computers: > And that's good. How many people know about the CHECKSUM command, either? Two years ago I invented it (that's MS speak, meaning I wrote a checksum program for the Winxx platforms when I had to catalog my eprom files ;-) It was done in such a crude way and was so slow BTW, that I won't even _give_ it away. I'd be too embarassed. If you know what a progress bar is, you probably also know what happens if you update one 65,536 times while computing the checksum of a 27512. If you fgetc the file a byte at a time on top op that, you can imagine the result. Positive side: the program was compiled 5 minutes after I started writing it. But maybe I aught to have spent 2 more minutes :) -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 21:23:13 GMT Organization: . Lines: 16 Message-ID: <36767d2d.5174310@news.uunet.be> References: <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> NNTP-Posting-Host: pool02b-194-7-79-161.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.tli.de!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach lucvdv@null.net (Luc Van der Veken) on Thu, 10 Dec 1998 23:49:16 GMT to alt.folklore.computers: > I've been asking myself serious questions about this approach, > ... > (haven't thought that far - never needed to). That sounds funny even to me, when I re-read it. What I meant is that the questions were of a serious nature, not that I gave them much thought. That clears that up. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 22:01:19 GMT Organization: Daedalus Consulting Lines: 40 Message-ID: <74s4nf$v2$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <74nqnu$1si$1@news.latrobe.edu.au> <74r7lo$l7g$1@news.wlg.netlink.net.nz> <74rnv6$62s@relay.lsl.co.uk> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74rnv6$62s@relay.lsl.co.uk>, Ben Hutchings wrote: >Don Stokes (don@news.daedalus.co.nz) wrote: > >: Later versions (post V5) play some funny games defining a "primary" and >: "secondary" links, defining the "primary" link as the one where the >It would allow a backup program, for example, to weed out extra links >to a file and only store one copy. (Of course, there are several >other ways to do this.) Nope. Backup figures this out by looking at file-IDs. If you do a BACKUP/IMAGE, the first thing it does is parse the entire index file (INDEXF.SYS). It does this for a couple of reasons -- (a) it can figure out where the links are, and (b) if you're backing up the whole disk it's a lot faster since it's not having to parse large directory files, going back to the index file for each file header -- makes an *enormous* difference if there are very large directories. (For some inadequately explained reason, directories stop being cached if the directory is over 100 or so blocks, and the performance loss from this is really noticable. I once had a news system where the expiry stuff failed, leaving me with some 30,000 files in one directory to get rid of by the time I noticed. DELETE *.*;* was deleting the files at an average rate of one every five seconds (it would speed up toward the end, but even so, the estimated time to completion was a day away). So I wrote a bit of DCL code to read through the index file and tweak the "delete pending" bit on all file headers with a backlink to the offending directory, then did a SET FILE/NODIR on the directory and deleted it. ANALYZE/DISK/REPAIR could then mop it all up nicely, deleting all the orphaned marked-for-delete files without having to reference a directory.) BACKUP without /IMAGE does not attempt to restore links. Which is a bloody nuisance for incremental backups of system disks on clusters, since all the cluster-common directories are linked below the node-specific directories (ie [SYSn.SYSCOMMON]), as well as to [VMS$COMMON]. So incremental backup command files almost always explicitly exclude [SYS%.SYSCOMMON], assuming the sysadmin has a clue. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 11 Dec 1998 22:47:39 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> NNTP-Posting-Host: d1.56.d0.27 X-Server-Date: 12 Dec 1998 04:47:40 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <74qpdt$57u@axalotl.demon.co.uk>, huge@nospam.demon.co.uk wrote: >So don't do it that way. What's wrong with > >chmod u+x,go-rwx filename I consider that even more obtuse than binary. Of course, I'm a programmer... -- Howard S Shubs The Denim Adept ###### From: ftit@engin.umich.edu (Sergej Roytman) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 11 Dec 1998 22:54:25 GMT Organization: University of Michigan, College of Engineering Lines: 9 Message-ID: <74s7r1$e61@srvr1.engin.umich.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74jpov$4iu$1@agate.berkeley.edu> <74k850$996$1@news.wlg.netlink.net.nz> <366E1D75.F5D3585F@tnglwood.demon.co.uk> NNTP-Posting-Host: az.engin.umich.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!141.211.144.13!newsxfer3.itd.umich.edu!news.eecs.umich.edu!srvr1.engin.umich.edu!ftit In article <366E1D75.F5D3585F@tnglwood.demon.co.uk>, Robert Billing wrote: > OTOH when I first moved from VMS to linux I kept wondering, "how do I >*stop* things iterating through all the drives?". find(1) has a -prune option for just that purpose. -- Sergej Roytman ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: lucvdv@null.net's message of Thu, 10 Dec 1998 23:49:16 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> Date: Fri, 11 Dec 1998 23:55:05 GMT Lines: 38 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!News.Amsterdam.UnisourceCS!masternews.telia.net!news-nyc.telia.net!newshub.northeast.verio.net!ix.netcom.com!netcom!alderson In article <367551c8.59800829@news.uunet.be> lucvdv@null.net (Luc Van der Veken) writes: >Also sprach hshubs@mindspring.com (Howard S Shubs) on Wed, 09 Dec >1998 22:51:59 -0600 to alt.folklore.computers: >> In article <74mabs$rrl$1@agate.berkeley.edu>, >> korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >> >> >Strange. Are you telling me that VMS directories weren't hacked onto a >> >flat file system? >> >> Can't say. They don't -look- like it. >Can they, in any way, be considered as files themselves? In that case you >could say they're hacked onto a flat file system, just like in every OS I >know. >Or is all information, including that needed to determine what directory an >entry belongs to, stored in one single giant bag? I can't answer for VMS, but Tops-20 directories are (a) hierarchical and (b) only file-like in that each sub-directory is represented by an entry in the parent directory. They cannot be opened as files by any system call, however. The File Descriptor Block for a directory file contains a pointer to the location on disk of the Directory Descriptor Block that contains all the pointers to the FDBs for that directory's files. There is a for each filesystem (which may reside on multiple disks), which even contains an entry for itself; the monitor finds it by looking in the HOME blocks (which are replicated on each disk of the filesystem) for the index page that points to it. (I think VMS is probably more like Tops-10, as much as I understand the Tops-10 filesystem.) -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 12 Dec 1998 03:44:40 GMT Organization: Daedalus Consulting Lines: 76 Message-ID: <74sor8$1so$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74qpkp$rm5$1@hades.csu.net> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <74qpkp$rm5$1@hades.csu.net>, who? wrote: >There seems to be a lot of confusion concerning *VMS logicals. To compare >them to symlinks is, at least in my mind, erroneous. They are more akin >to global variables, if you were to consider shell variables and DCL >symbols to be local variables. A bit like perl's hash arrays, they allow >programs to look up a value given a name. DCL symbols can be got at by applications. Check out LIB$GET_SYMBOL and LIB$SET_SYMBOL -- I'd call those rather better analogues to environemt variables, although the match is by no means perfect. (For example, if you set a symbol, it will be accessible to DCL and following commands after the current image terminates.) And symbols don't really act like hash arrays, although because DCL is interpreted from source on every line, you can play games like: $ foo_a = "hello" $ foo_b = "world" $ i = "a" $ write SYS$OUTPUT: foo_'i' hello $ i = "b" $ write SYS$OUTPUT: foo_'i' world >As it happens, they are frequently used to refer to parts of the file >system, but there is no reason why they can't contain a network address, >or a terminal device. Like shell vars, it's up to the program to decide >what they represent. There's no reason why a symlink can't contain a network address or a terminal device. Why is: % ln -s /dev/tty00 foo % ln -s 192.168.1.1 my_address % ln -s "Hello, World" message signficantly different from $ assign TXA0: foo $ assign 192.168.1.1 my_address $ assign "Hello, World" message I can read the symlink with readlink(2), and the logical with $TRNLNM. What's important about both logicals and symlinks, and doesn't occur with DCL symbols or environemt variables, is that the file system reconises these without program intervention -- I don't need to use $TRNLNM to access a file via a logical just as I don't have to call getlink() to follow a symlink. Where the two differ is that logicals live in a different namespace from directories and files, whereas symlinks live in one common namespace. A single namespace is the way unix does things. Separate namespaces for directories, files, devices and so-on is the way VMS does things, and it's quite a bit more complicated because of that. On the other hand separate namespaces mean that namespaces can have different and more appropriate individual characteristics, whereas with unix you're stuck with a single, common, not particularly dynamic namespace -- eg unix devices aren't accessible via the filesystem unless someone has explicitly created a suitable filesystem entry for it (usually in /dev), whereas VMS device names are created and destroyed on the fly along with the actual incarnation of the device. Similarly, logicals logicals can apply to individual processes or groups rather than being systemwide. Logicals, living in a somewhat ephemeral namespace, do go away whereas symlinks stick around until they're explicitly deleted, but that's just the nature of those namespaces. VMS folks just know if they want a logical to stick around longer than the current image, process, job or system incarnation, they have to stick the assignments in the appropriate startup/login/whatever command file. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jkatz@ultrix6.cs.csubak.edu (who?) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 12 Dec 1998 08:33:18 GMT Organization: California State University, Bakersfield Lines: 70 Message-ID: <74t9oe$35g$1@hades.csu.net> References: <73ela0$51h@hermes.acs.unt.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74qpkp$rm5$1@hades.csu.net> <74sor8$1so$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: ultrix6.cs.csubak.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!usc!newshub.csu.net!ultrix6.cs.csubak.edu!jkatz Don Stokes (don@news.daedalus.co.nz) wrote: : DCL symbols can be got at by applications. Check out LIB$GET_SYMBOL and : LIB$SET_SYMBOL -- I'd call those rather better analogues to environemt : variables, although the match is by no means perfect. (For example, if : you set a symbol, it will be accessible to DCL and following commands : after the current image terminates.) If you define them to last beyond the current process, a matter of an extra '=' in the decleration when defined via DCL. : And symbols don't really act like hash arrays, although because DCL is : interpreted from source on every line, you can play games like: Ahh, you reversed what I meant. Logicals in my opinion are more like hash arrays, although they have the curious ability to have multiple values (I think sys$common is this way). One of the really nice things about logicals that I had forgotten earlier, that both symbols and environment variables lack, is the ability to affect changes accross the system. If you change sys$common (for some reason..) in the system table, every user will see the change the next time the try to access it, not the next time they log in, or source the system wide login procedure. While admitedly, you could do this with symlinks, they don't strike me as being as flexible. Not as robust either. : Why is: [text deleted for easier reading] : $ assign "Hello, World" message And I learn something new about *NIX environments every day. I think this qualifies as the most interesting tidbit today. However, they are tied to the file system, as if you were to unmount the disk containing them, they "go away". Logicals association with the file system is completely at the descretion of the user assigning them. Unless of course the page file happens to be on the suddenly unavailable disk. And yet, just now, I tried: ln -s 127.0.0.1 local; telnet ./local; and was greeted with "telnet: no such host". Symlinks don't care whether the target exists or not, but trying to use them for something that is not a file runs accross problems unless the program using them goes to an extra effort to distinguish them as something else, Im guessing via lstat() or something of the like. : Where the two differ is that logicals live in a different namespace from : directories and files, whereas symlinks live in one common namespace. A : single namespace is the way unix does things. Separate namespaces for : directories, files, devices and so-on is the way VMS does things, and : it's quite a bit more complicated because of that. and yet I can still do $ open blackboard: opa0: $ write blackboard: "teach me something" $ close blackboard: Just because I cant delete it, doesnt mean that I cant do useful things with very little consideration as to the fact that it is not a file. : Logicals, living in a somewhat ephemeral namespace, do go away whereas : symlinks stick around until they're explicitly deleted, but that's just : the nature of those namespaces. VMS folks just know if they want a : logical to stick around longer than the current image, process, job or : system incarnation, they have to stick the assignments in the : appropriate startup/login/whatever command file. Not quite. You just have to stick it in the appropriate table. Of course, your average user (TMPMBX, NETMBX) doesnt have the ability to write to anything more permanent than their job table. Symbols are of the more ephemeral space, being limited to a process, and optionally, is children (which doesnt seem quite the right word). jeremy ###### From: huge@nospam.demon.co.uk (Hugh Davies) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 12 Dec 1998 09:22:29 GMT Organization: Piglet's Pickles and Preserves Message-ID: <74tckl$al@axalotl.demon.co.uk> References: Reply-To: huge@nospam.demon.co.uk NNTP-Posting-Host: axalotl.demon.co.uk X-NNTP-Posting-Host: axalotl.demon.co.uk:158.152.24.143 X-Trace: news.demon.co.uk 913456841 nnrp-01:7447 NO-IDENT axalotl.demon.co.uk:158.152.24.143 X-Complaints-To: abuse@demon.net x-no-archive: yes Lines: 22 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news.demon.co.uk!demon!axalotl.demon.co.uk!axalotl!usenet In article , hshubs@mindspring.com (Howard S Shubs) writes: >In article <74qpdt$57u@axalotl.demon.co.uk>, huge@nospam.demon.co.uk wrote: > >>So don't do it that way. What's wrong with >> >>chmod u+x,go-rwx filename > >I consider that even more obtuse than binary. Really? I find it so much easier that I've never bothered to learn the bitwise interpretation. > Of course, I'm a programmer... I *used* to be. Never on Unix, tho' -- "The road to Paradise is through Intercourse." The uk.transport FAQ; http://www.axalotl.demon.co.uk/transport/FAQ.html [Substitute "axalotl" for "nospam" to email me] ###### From: David Wragg Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Lines: 23 Message-ID: References: <74mmn9$4l7$5@irk.zetnet.co.uk> Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Date: 12 Dec 1998 00:32:33 +0000 NNTP-Posting-Host: 194.119.176.228 X-Complaints-To: news@u-net.net X-Trace: newsr2.u-net.net 913511887 194.119.176.228 (Sun, 13 Dec 1998 01:18:07 BST) NNTP-Posting-Date: Sun, 13 Dec 1998 01:18:07 BST Organization: (Posted via) U-NET Internet Ltd. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!194.72.7.126!btnet-peer!btnet!peer.news.th.u-net.net!u-net!newsr2.u-net.net.POSTED!gatsby.u-net.com!not-for-mail lisard@zetnet.co.uk writes: > Unix, on the other hand, *does* do globbing in the application. The > shell is the application that does the globbing; Well, $ nm /lib/libc.so.6 | grep -w glob 0005e650 T glob Libc is part of the OS by any reasonable standard. > the idea is that > nothing else then has to worry about it. On the other hand, since a > directory is merely a list of filenames and i-nodes, you can open a > directory as a file, which eases the pain somewhat. This is correct for Unix in the historical sanse. However, POSIX says open() can return EISDIR, and specifies opendir() et al (as does BSD, SVID), and most systems do so. Any program today that tries opening directoris and reading 2 byte inode number, 14 byte filename records won't be very portable. Dave Wragg ###### From: David Wragg Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> <1998Dec8.101907.3786@lorelei.approve.se> Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Date: 12 Dec 1998 01:46:01 +0000 NNTP-Posting-Host: 194.119.176.228 X-Complaints-To: news@u-net.net X-Trace: newsr2.u-net.net 913511888 194.119.176.228 (Sun, 13 Dec 1998 01:18:08 BST) NNTP-Posting-Date: Sun, 13 Dec 1998 01:18:08 BST Organization: (Posted via) U-NET Internet Ltd. Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!dispose.news.demon.net!demon!peer.news.zetnet.net!peer.news.bb.u-net.net!u-net!newsr2.u-net.net.POSTED!gatsby.u-net.com!not-for-mail meowing@banet.net (Fluffy) writes: > Hmm. If Unix style shell globbing is a Good Idea, why do I so often > find myself having to rely on ugly hacks like xargs to run things > against large sets of files? You have it the wrong way round. Filename wildcarding is an ugly (though useful) hack. Find and xargs are elegant bridges between some of the "domains" of Unix: - Filesystem data (directory contents and file meta data) - Text streams - Command lines The other main domain is file contents. Other such bridges are redirection, cat and echo. "Your aeroplane is so noisy". "But your bicycle is so slow". Aren't religious wars fun? Dave Wragg ###### Message-ID: <36724E3F.449CDFA3@trailing-edge.com> From: Tim Shoppa Organization: Trailing Edge Technology X-Mailer: Mozilla 3.03Gold (X11; I; OpenVMS V7.0 DEC 3000 Model 300L) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <74u0qq$4kf$4@ligarius.ultra.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 11 Date: Sat, 12 Dec 1998 11:06:39 -0400 NNTP-Posting-Host: 198.232.144.27 X-Trace: audrey2.cais.com 913479246 198.232.144.27 (Sat, 12 Dec 1998 11:14:06 EDT) NNTP-Posting-Date: Sat, 12 Dec 1998 11:14:06 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!nntp.abs.net!outfeed1.news.cais.net!in1.nntp.cais.net!199.0.216.204.MISMATCH!audrey2.cais.com!not-for-mail jmfbahciv@aol.com wrote: > What the hell does flat file system mean and why is it > obnoxious to some people? Good question, and I wish I knew the answer. As I pointed out previously in this thread, most all computers that ordinary folks have access to treat the disk as a collection of numerically- addressed blocks; at this level all file systems (including VMS and Unix) are "flat". Tim. ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 1998 12:09:11 +0100 Organization: [Posted via] Interactive Networx Lines: 14 Message-ID: <1djw212.149to71o4xrr4N@n164-76.berlin.snafu.de> References: <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <1djvj1d.8wyafb10bgojyN@n244-87.berlin.snafu.de> <74r7qj$rbj$1@pegasus.csx.cam.ac.uk> NNTP-Posting-Host: n164-76.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news.gsl.net!gip.net!newsfeed.tli.de!unlisys!news.snafu.de!jnickelsen Ben Harris wrote: [Unix files] > > no creation date. > > Erm, st_ctime? st_ctime is the time of the last status (i. e. inode) change. Other than last modification time and last access time, which can be changed via the utimes(2) system call, it cannot be changed by the user, but rather is updated when one of the other file attributes is changed. -- Juergen Nickelsen ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 98 13:57:01 GMT Organization: UltraNet Communications, Inc. Lines: 83 Message-ID: <74tu3k$4kf$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> NNTP-Posting-Host: d5.dial-16.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 12 Dec 1998 14:20:36 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d5 In article , alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >In article <74oi4r$nrp$4@ligarius.ultra.net> jmfbahciv@aol.com writes: > >>>AIX, like some other modern Unix implementations, doesn't >>>have a separate >>>pager process. All disk I/O (except I/O to special partitions that are >>>reserved to special third-party applications, like DBMSes, and largely >>>ignored by the OS) goes through the kernel's virtual memory >>>manager. That >>>means all swap and file I/O is memory-mapped, under the covers ("covers" >>>being the read, write, etc. system calls) if necessary. > >>OK, so I'm assuming that this VM manager is a bottleneck. >>And that doesn't sound very good for performance (understand, >>that I come from a group who was >>terribly fussy about good performance w.r.t the user). It >>also means that >>general I/O is funneled through the VM code. That just gives me the >>shivers--of course, it could be that I'm biased to the hilt :-). > >Could be, Barb. ;-) Highly likely :-). > >After all, that's the I/O model used in Tops-20, and its >predecessor TENEX, hardly "modern" in the sense the youngsters >mean it. Yup. And that model was terribly slow compared to the system I worked with. > >Oh, by the way, the I/O performance was good on DEC disks, and is still good on >SCSI-2 disks from an old DEC customer. (How many people these days know that >Hewlett-Packard used a Stanford-originated CAD package that ran on a DEC-20?) > >>>The VMM, which is a kernel lightweight process, won't block >>>forever on I/O >>>problems because it has a watchdog timer. > >>Good gads, it shouldn't! The requestor should probably have >>the timeout for retries or whatever is appropriate for that >>particular device medium (note >>that this should be at the device driver level). > >I think some confusion has set in: The VM code in this case >*is* the requestor for services from an appropriate disk driver, >and will time out eventually if the driver does not come back to it. Well, we're getting into the nitty gritty that I wasn't familiar with in my line (that was for JMF and TW to do). It doesn't seem to me that the request for the disk driver should block until the driver finished its I/O. The requestor should be able to be notified when the request is done, freeing it up for other work. > >>>With the all-I/O-via-VMM architecture, it's not a problem. The OS knows >>>which pages are in RAM and which are not. > >>It is a problem if you wanted to do symmetric multi-processing. >>It doesn't >>seem to allow I/O that is synchronous with the user requests but uses an >>asynchonous process. (Or, I'm simply not understanding what you're >>describing.) > >That may be why Tops-20 can never be made SMP; I've had it >explained to be in the past, but since I was only ragging >on someone about fiddles in the OS to start with, I didn't >try to remember all the details of the explanation. :-) Well, JMF figured out how to SMP TOPS-20 but it was too late when he presented it. I've been told that these were also problems when he did SMP UNIX. What the guy couldn't remember is if they ever solved the problem of contention. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 98 14:31:46 GMT Organization: UltraNet Communications, Inc. Lines: 193 Message-ID: <74u04s$4kf$3@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> <74p9fu$a0n@news3.newsguy.com> NNTP-Posting-Host: d5.dial-16.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 12 Dec 1998 14:55:24 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.244.253.199!newsfeed.xcom.net!news.ultranet.com!d5 In article <74p9fu$a0n@news3.newsguy.com>, mww@microfocus.com (Michael Wojcik) wrote: > >In article <74oi4r$nrp$4@ligarius.ultra.net>, jmfbahciv@aol.com writes: >> In article <74kkr4$6t9@news1.newsguy.com>, >> mww@microfocus.com (Michael Wojcik) wrote: >> > >> >In article <74glh7$j1n$1@strato.ultra.net>, jmfbahciv@aol.com writes: >> >> [somebody's hacked off the atributes and it tweren't me] >> >> >> >> In article <366b39eb.3006571501@news.cyberport.com>, >> >> tangentSPAMCATCHER@cyberport.com (Warren Young) wrote: >> > >> >> >I've only been working with computers seriously since about 1990 (and >> >> >a nerdy little whiz kid since 1985), but it seems to me that you could >> >> >do this with a preemptible, multitasking kernel. [...] >> >> >> >> You must consider the case where your microkernel preempts itself. >> > >> >The AIX kernel, as of v3, is preemptible as well as pageable. And >> >it's not even a microkernel. It's not a trivial design, but it's >> >not prohibitively difficult. Mostly it means putting a lot of >> >fast synchronization mechanisms in at the right places. >> >> But, if I understand the implementation, is it re-entrant? > >Some parts of it are, at least as I understand it. It was designed to >support real-time processes (though I'll let others decide to what >extent that effort was successful), and so the kernel is preemptible, >pageable, reentrant, and supports fixed absolute process priorities. >It has fast synchronization mechanisms, since obviously there are a >lot of critical sections in a full Unix-style kernel. Right. There was a unique project in the TOPS-10 SMP implementation to identify those areas and keep the instruction set (ohh, that's the wrong term...it'll be misunderstood) to the minimum. IIRC, the instruction window was two. We had a lingo for this but I sure as hell can't remember it right now. > >That's why, for example, the "cs" (compare and swap) atomic system >call provided by the AIX 3 kernel used SVC 2 (supervisor call 2) rather >than SVC 1, as all other system calls did; SVC 1 calls could preempt >one another and reenter. SVC 2 *wasn't* preemptible. (cs was dropped >in AIX 4, because it was too difficult to get it working nicely on >multiprocessor machines. It was a hack, albeit a relatively neat one.) Sometimes that's all one can do :-). > >My main reference for this (besides a bunch of AIX internals discussions >I sat in on when I was at IBM) is "Enhancements to the AIX Kernel for >Support of Real-Time Applications", by Kathy Bohrer and John O'Quin, >in _IBM RISC System/6000 Technology_, an IBM pub that collected several >RS/6000 and AIX papers from IBM journals. This may be the cause of some of our misunderstandings. Real time is very different from timesharing and the approaches required for real time won't work for timesharing...and visa versa. It's just the nature of the beast. > >Note that interrupt handlers, while preemptible (by kernel processes >at higher priority), are not pageable and presumably not reentrant. But interrupts are done on a per CPU basis. So that doesn't make any sense. There definitely will be code that can't be interrupted until the data necessary to handle the interrupt is stored in a safe place for the code that will take care of the event that caused the interrupt. But that's a very, very small code window (or should be for performance's sake, anyway). >As is generally the case in Unix, a poorly-written interrupt handler >can hang or crash the system. Well, of course :-). > Also, most of the synchronization >primitives in the kernel are cooperative, at least in the AIX 3 design. >(I don't know if AIX 4 enforces them.) There are plenty of potential >pitfalls in this sort of implementation, but it's a start. > >> >AIX, like some other modern Unix implementations, doesn't have a >> >separate pager process. All disk I/O (except I/O to special >> >partitions that are reserved to special third-party applications, >> >like DBMSes, and largely ignored by the OS) goes through the kernel's >> >virtual memory manager. That means all swap and file I/O is >> >memory-mapped, under the covers ("covers" being the read, write, etc. >> >system calls) if necessary. >> >> OK, so I'm assuming that this VM manager is a bottleneck. And that >> doesn't sound very good for performance (understand, that I come >> from a group who was terribly fussy about good performance w.r.t >> the user). It also means that general I/O is funneled through >> the VM code. That just gives me the shivers--of course, it could >> be that I'm biased to the hilt :-). > >Performance is fine, unless you run out of virtual memory; but >thrashing is thrashing, and AIX does have a mechanism for recovering >more or less gracefully from memory exhaustion (it starts killing >off processes in some "least-favored" order, first with a catchable >signal so the process can exit cleanly and then with an uncatchable >one to enforce it, until it's below the danger mark). One of the things that I've been forgetting is that hardware nowadays isn't as slow as in my day. So the bottlenecks move to the slowest as hardware is upgraded. This was also true when TW would do the device drivers for new hardware. Putting a faster disk on a system abruptly caused the same user mix to become CPU bound. Putting on a faster CPU would invariably cause the same mix to become I/O bound. After a few experiences we could predict the effect of a new piece of hardware and plan a mini-project for performance improvements in the areas of the monitor that had nothing to do with the new hardware. And, every once in a while, JMF or TW would be completely surprised by a new side effect; that's when they had fun. > >The VMM controls all the RAM and disk storage. It deals with pages, >and it manages them using the hardware TLBs and inverse TLBs. When >a process wants a page, chances are that it's already available in >RAM, because RAM is effectively an MRU cache of pages. If it's not, >the process gets preempted while the VMM schedules a page steal and >page-in. If a process writes to a file, that page (or those pages) >are marked for commitment (they'll be committed when someone issues >a sync against that file or against the whole system; that happens >periodically in the background if no user process does it explicitly). >If another process tries to read the same part of that file, they'll >get the copy from memory - the VMM knows that that file page is in >RAM already. > >The VMM isn't generally a bottleneck because generally the most >frequently accessed parts of the combined working sets of running >processes are in RAM already. I have a number of Unix workstations >I use daily (I develop and maintain a Micro Focus product on several >Unix flavors), and my RS/6000s pause least frequently for disk I/O, >be it file or swap. They don't have more RAM than my other machines; >they just balance the scheduling better. How many jobs typically run on one of those machines? > >Also, the kernel supports pinning pages in RAM, so the most frequently >used kernel tables and whatnot never get paged out. A useful aspect. Put in the wrong hands.... :-) we had customers like that. >> >Typically the kernel dumps core to >> >the save device, if it can, and shuts down or reboots. Many modern >> >machines have some facility for storing some error information in >> >non-volatile memory, too. >> >> TOPS10 also had a "dump core and continue" stopcode category. That >> was the basis for implementing the "suspend" option of Symmetric >> Multi-Processing. > >The problem with Unix is that the panic may have been issued by >third-party device driver code or kernel extensions. Who's to say >what's been damaged? The damage detector. > The typical Unix kernel does not provide a >lot of protection against kernel code going haywire. There are tricks. But coders of kernal software should know what they are doing. We were able to have the luxury of having pseudo-proprietary software. We did not have to worry about customers munging in things they didn't know anything about ,although we did give them all the opportunity since we shipped our sources. Customer-developed code was a constant give and take. Some of it was good, a lot of it was bad. And a lot of our art was picking which we would look at. That was the business end of our work. I'm discovering that people who do development don't realize that coding is just 1% of the effort (even though it may take 50% of ones time). > >Fault-tolerant Unix "systems" are generally clusters of semi-independent >machines that monitor for one another and do takeovers when necessary. Yup. JMF never did decide which was the better implmentation...a fault-tolerant or SMP. Each had their advantages and each had their disadvantages. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 98 14:43:32 GMT Organization: UltraNet Communications, Inc. Lines: 60 Message-ID: <74u0qq$4kf$4@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> NNTP-Posting-Host: d5.dial-16.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 12 Dec 1998 15:07:06 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!news.maxwell.syr.edu!newsfeed.xcom.net!news.ultranet.com!d5 In article , alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >In article <367551c8.59800829@news.uunet.be> lucvdv@null.net >(Luc Van der Veken) writes: > >>Also sprach hshubs@mindspring.com (Howard S Shubs) on Wed, 09 Dec >>1998 22:51:59 -0600 to alt.folklore.computers: > >>> In article <74mabs$rrl$1@agate.berkeley.edu>, >>> korpela@islay.ssl.berkeley.edu (Eric J. Korpela) wrote: >>> >>> >Strange. Are you telling me that VMS directories weren't hacked onto a >>> >flat file system? >>> >>> Can't say. They don't -look- like it. > >>Can they, in any way, be considered as files themselves? >>In that case you could say they're hacked onto a flat file >>system, just like in every OS I know. > >>Or is all information, including that needed to determine >>what directory an entry belongs to, stored in one single giant bag? > >I can't answer for VMS, but Tops-20 directories are (a) hierarchical >and (b) only file-like in that each sub-directory is represented by >an entry in the parent directory. They cannot be opened as files >by any system call, however. The File Descriptor Block for a >directory file contains a pointer to the location on disk of >the Directory Descriptor Block that contains all the >pointers to the FDBs for that directory's files. There is >a for each filesystem (which may reside on >multiple disks), which even contains an entry for itself; the >monitor finds it by looking in the HOME blocks (which >are replicated on each disk of the filesystem) for the index >page that points to it. > >(I think VMS is probably more like Tops-10, as much as I >understand the Tops-10 filesystem.) I don't know the VMS file system structure so I can't say [emoticon wondering if its just been insulted a tad :-)]. The 20's is equivalent to the -10's MFD that had the PPN of [1,1]. As a matter of fact, in the Level C disk architecture, one could actually log in as [1,1] and do nasty things like DELETE *.*. An operator at the university installation forgot s/he was logged in as [1,1], had been doing other work as if s/he was logged in to her/his directory, decided to clean up, and started to wipe out the whole file system. That was "fixed" with the implementation of Level D. What the hell does flat file system mean and why is it obnoxious to some people? /BAH Subtract a hundred and four for e-mail. ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 1998 16:28:55 GMT Organization: . Lines: 14 Message-ID: <36728afd.10991534@news.uunet.be> References: <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> NNTP-Posting-Host: pool02b-194-7-80-206.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach alderson@netcom.netcom.com (Richard M. Alderson III) on Fri, 11 Dec 1998 20:03:04 GMT to alt.folklore.computers: > Unless Linux is very different from the Unices I've worked with (SunOS, HP-UX, > Solaris, Ultrix, OSF/1, Digital Unix 4.0, Pyramid), you can't copy files to an > unmounted filesystem, so this can't work. You can temporarily mount it on another mount point just for copying. How else would you - when setting it up - get all the existing files in there in the first place? -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 1998 17:54:44 -0600 Organization: MindSpring Enterprises Lines: 14 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <74u0qq$4kf$4@ligarius.ultra.net> <36724E3F.449CDFA3@trailing-edge.com> NNTP-Posting-Host: d1.56.d0.13 X-Server-Date: 12 Dec 1998 23:51:25 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <36724E3F.449CDFA3@trailing-edge.com>, Tim Shoppa wrote: >Good question, and I wish I knew the answer. As I pointed out >previously in this thread, most all computers that ordinary folks >have access to treat the disk as a collection of numerically- >addressed blocks; at this level all file systems (including >VMS and Unix) are "flat". Now I see what you're saying. At the level you're talking about, there are no files; that's below the file system at the physical disk level. At that level, only blocks exist. That's the way disks work. -- Howard S Shubs The Denim Adept ###### Message-ID: <3672C59F.7D85EFD8@trailing-edge.com> From: Tim Shoppa Organization: Trailing Edge Technology X-Mailer: Mozilla 3.03Gold (X11; I; OpenVMS V7.0 DEC 3000 Model 300L) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <74u0qq$4kf$4@ligarius.ultra.net> <36724E3F.449CDFA3@trailing-edge.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 26 Date: Sat, 12 Dec 1998 19:35:59 -0400 NNTP-Posting-Host: 198.232.144.27 X-Trace: audrey2.cais.com 913509809 198.232.144.27 (Sat, 12 Dec 1998 19:43:29 EDT) NNTP-Posting-Date: Sat, 12 Dec 1998 19:43:29 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!in1.nntp.cais.net!199.0.216.204.MISMATCH!audrey2.cais.com!not-for-mail Howard S Shubs wrote: > > In article <36724E3F.449CDFA3@trailing-edge.com>, Tim Shoppa > wrote: > > >Good question, and I wish I knew the answer. As I pointed out > >previously in this thread, most all computers that ordinary folks > >have access to treat the disk as a collection of numerically- > >addressed blocks; at this level all file systems (including > >VMS and Unix) are "flat". > > Now I see what you're saying. At the level you're talking about, there > are no files; that's below the file system at the physical disk level. At > that level, only blocks exist. That's the way disks work. Well, not the way that all disks work (there do exist associative storage units, many of which don't even have fixed-sized blocks and instead resector the disk surface on the fly), but it's the level that VMS and Unix starts with. Note that some languages/operating systems treat the numbered blocks on disk as the file system and don't build anything on it. The original Forth is the obvious example. (Though some will argue that Forth is neither a language nor an OS...) Tim. ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 1998 22:41:13 -0600 Organization: MindSpring Enterprises Lines: 33 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <74u0qq$4kf$4@ligarius.ultra.net> <36724E3F.449CDFA3@trailing-edge.com> <3672C59F.7D85EFD8@trailing-edge.com> NNTP-Posting-Host: d1.56.d2.0b X-Server-Date: 13 Dec 1998 04:41:14 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <3672C59F.7D85EFD8@trailing-edge.com>, Tim Shoppa wrote: >Well, not the way that all disks work (there do exist associative >storage units, many of which don't even have fixed-sized blocks >and instead resector the disk surface on the fly), but it's the >level that VMS and Unix starts with. It's the way standard disks work. Associative storage devices... well, I've not heard of any outside of academia but I'll take your word that they exist. Still, every OS I've ever heard of (granted, I don't go back as far as some here) accesses disks as block-structured devices. Floppies and "winchester" disks... blocks. VMS, UNIX, Macintosh, Microsot/IBM: blocks. Certainly it's possible to create something which does it differently; this is hardware/software, it can be designed to do just about anything a designer can come up with. But the most useful, to date, disk devices are block structured. If I'm wrong, I'm really interested in the correction! And this is -the- crowd to correct me, too. Disk-based file systems start with devices containing blocks. Then you could bring up things like TU58 tapes formatted as Files-11, but please don't. >Note that some languages/operating systems treat the numbered >blocks on disk as the file system and don't build anything on it. >The original Forth is the obvious example. (Though some will >argue that Forth is neither a language nor an OS...) There's always the possability of some specialized system doing things radically different, yes. I would even enjoy such stuff. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 12 Dec 1998 22:47:29 -0600 Organization: MindSpring Enterprises Lines: 23 Message-ID: References: <74tckl$al@axalotl.demon.co.uk> <74v2ap$pbc$1@news1.bu.edu> NNTP-Posting-Host: d1.56.d2.0b X-Server-Date: 13 Dec 1998 04:47:29 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <74v2ap$pbc$1@news1.bu.edu>, slanning@buphy.bu.edu (Scott Lanning) wrote: >For permission to read OR write, bitwise OR > > r = 100 > w = 010 >--------------- (mnemonic: ones are heavy, > r | w = 110 so they drop through) > >At this point, add from left to right 4 + 2 + 0 = 6. For that, >you must know binary....only to three places. Yes, that's the level of binary I taught 'em. Only three bits were necessary, repeated twice. I use the symbolic method myself when changing across the board, as in chmod +w foo to make file foo writable to everyone, but I find the qualifiers for u-ser, g-roup, and o-ther to be kinda tough to combine with the rest, so I tried for simplicity. I don't think it sunk in, and I doubt I'll try that method again. Maybe yours would have been less (seemingly) overwhelming. Think I'll print out your message and give them copies, if you don't mind. -- Howard S Shubs The Denim Adept ###### From: slanning@buphy.bu.edu (Scott Lanning) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 13 Dec 1998 00:38:49 GMT Organization: Boston Univ. Physics Lines: 61 Message-ID: <74v2ap$pbc$1@news1.bu.edu> References: <74tckl$al@axalotl.demon.co.uk> NNTP-Posting-Host: buphy.bu.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!news.eecs.umich.edu!news.bu.edu!slanning huge@nospam.demon.co.uk(Hugh Davies): :hshubs@mindspring.com(Howard S Shubs): : >I consider that even more obtuse than binary. : :Really? I find it so much easier that I've never bothered to learn :the bitwise interpretation. I don't find it easier, but I think the symbolic method supersedes the absolute because you can set permissions absolutely with symbols (go=rx), whereas you can't set permissions relatively with the absolute method (can't do go-r, you have to choose 611 or whatever). Initially, at least, I tend to think relatively, e.g. (chmod go-wx foo.bar), ? Group AND Other ? CANNOT ? Write OR eXecute ? foo.bar rather than absolutely, e.g. (chmod 644 foo.bar), ? User ? CAN Read OR Write ? CANNOT eXecute ? AND ? Group ? CAN Read ? CANNOT Write OR eXecute ? AND ? Other ? CAN Read ? CANNOT Write OR eXecute ? foo.bar As for learning the bitwise interpretation, it's not too hard. read = 100 (mnemonic: Read, Write, eXecute write = 010 is ordered alphabetically). execute = 001 For permission to read OR write, bitwise OR r = 100 w = 010 --------------- (mnemonic: ones are heavy, r | w = 110 so they drop through) At this point, add from left to right 4 + 2 + 0 = 6. For that, you must know binary....only to three places. Or, memorize every permutation (likely after you use them a while: 7=all,6=rw, 5=rx,4=r, and 0=none are all you usually need). Next step. User, Group, Other are ordered left to right, so User Group Other (mnemonic: I'm more important than others) ---- ----- ----- 6 4 4 would let user read and write, while group and other read-only. One can further modify executability (symbolically with X,s,t,l or absolutely by prepending an octal digit), but that will be left as an exercise to the reader.. --Scott :) ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 13 Dec 1998 10:54:51 GMT Organization: . Lines: 25 Message-ID: <367a92b5.1515509@news.uunet.be> References: <74tckl$al@axalotl.demon.co.uk> <74v2ap$pbc$1@news1.bu.edu> NNTP-Posting-Host: pool02b-194-7-80-118.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach slanning@buphy.bu.edu (Scott Lanning) on 13 Dec 1998 00:38:49 GMT to alt.folklore.computers: > For permission to read OR write, bitwise OR > > r = 100 > w = 010 > --------------- (mnemonic: ones are heavy, > r | w = 110 so they drop through) Not bad if you're explaining it to a programmer, but I think for 'ordinary' people the confusion starts at the first line. In common speach 'OR' usually (or at least often) means exclusive or - [either] this OR that - while they would see what you were doing as AND - the user that gets this permission can [both] read AND write. "g-wx" for example would be read as "the roup can <->not rite nor eecute" - all they have to remember is that the <-> applies to all that follows it, not just the first character. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 13 Dec 98 12:23:19 GMT Organization: UltraNet Communications, Inc. Lines: 53 Message-ID: <750d08$anm$3@ligarius.ultra.net> References: <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <74u0qq$4kf$4@ligarius.ultra.net> <36724E3F.449CDFA3@trailing-edge.com> <3672C59F.7D85EFD8@trailing-edge.com> NNTP-Posting-Host: d11.dial-15.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 13 Dec 1998 12:47:04 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!18.24.4.11!newsswitch.lcs.mit.edu!news.ultranet.com!d11 In article , hshubs@mindspring.com (Howard S Shubs) wrote: >In article <3672C59F.7D85EFD8@trailing-edge.com>, Tim Shoppa > wrote: > >>Well, not the way that all disks work (there do exist associative >>storage units, many of which don't even have fixed-sized blocks >>and instead resector the disk surface on the fly), but it's the >>level that VMS and Unix starts with. > >It's the way standard disks work. No. It's the way the OS defines the geometry of the disk w.r.t. laying down the data. > Associative storage devices... well, >I've not heard of any outside of academia but I'll take your word that >they exist. Still, every OS I've ever heard of (granted, I don't go back >as far as some here) accesses disks as block-structured devices. Floppies >and "winchester" disks... blocks. VMS, UNIX, Macintosh, Microsot/IBM: >blocks. Certainly it's possible to create something which does it >differently; this is hardware/software, it can be designed to do just >about anything a designer can come up with. But the most useful, to date, >disk devices are block structured. If I'm wrong, I'm really interested in >the correction! And this is -the- crowd to correct me, too. I'm not up on the bitty bytey of establishing a disk geometry. You might wish to look at a device service routine that is written for disks that don't have a microprocessor in the controller (do any exist today?). Look at old listings of TOPS10 and compare the way the hardware was read. A really good comparison would be the RP0x device drivers and the DECtape device driver. Hint: latency had a lot to do with the establishment of the geometry. > >Disk-based file systems start with devices containing blocks. Then you >could bring up things like TU58 tapes formatted as Files-11, but please >don't. Well, I brought up TU58s with TOPS-10 format but an -11 that had disk and DECtape support is a valid comparison, too. DECtapes were nifty (and my favorite) storage devices but data was written to them in a surprising different way than disks. Just because they had blocks on them doesn't mean that the geometry of the actual 1s and 0s were the same. And, I suspect that's where you're confused. > > /BAH Subtract a hundred and four for e-mail. ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 13 Dec 1998 18:43:38 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 23 Message-ID: <36740a40.27720783@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!195.252.142.12!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On Sun, 06 Dec 1998 23:26:27 GMT, dg@ (David Given) wrote: >Part of the reason is the particularly lousy PC I/O protocols (I assume >you're using IDE here, SCSI's rather better). What *should* happen is that >when you take the hard-drive off-line, any processes with I/O transactions >pending (including swap) should block and everything else should keep >running. When you plug it back in the transactions should complete and the >processes unblock. This requires a reasonably secure I/O protocol between >the computer and the drive. I believe Amiga hardware requires smart >drives, which would explain why that still ran. I have had some experience with that on Amiga hardware. The commercial port of System V Release 4.0 to the Amiga exhibits exactly the behaviour you mention: when the whole of the SCSI bus goes AWOL (complete lock-up), every process continues running until the next disk access, at which time it just freezes solid. Unfortunately, the only way to un-lock the SCSI bus is to turn the whole machine off ... -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: Frank McConnell Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 13 Dec 1998 20:14:41 -0800 Organization: Reanimators Lines: 11 Message-ID: <7523bh$cm$1@daemonweed.reanimators.org> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> NNTP-Posting-Host: daemonweed.reanimators.org Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news.stanford.edu!news.kjsl.com!news.reanimators.org!not-for-mail huge@nospam.demon.co.uk (Hugh Davies) wrote: > So don't do it that way. What's wrong with > > chmod u+x,go-rwx filename It confuses folks like me who have had enough exposure to VMS that we have to sit there and think about whether "o" means "owner" like it does in SET FILE/PROT. Generally speaking, by the time I remember I worked out what I want in octal anyway. -Frank McConnell ###### From: pechter@news.monmouth.com (Bill Pechter) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 13 Dec 1998 21:06:34 -0500 Organization: Lakewood MicroSystems, Tinton Falls, NJ. Lines: 50 Message-ID: <751rra$11i4@i4got.pechter.dyn.ml.org> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> NNTP-Posting-Host: bg-tc-ppp24.monmouth.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newspeer.monmouth.com!news.monmouth.com!not-for-mail In article , Richard M. Alderson III wrote: >In article <1djrsod.6w627whz8fvgN@[10.0.0.3]> jnickelsen@acm.org >(Juergen Nickelsen) writes: > >>Mounting a file system over a non-empty directory just isn't done usually, >>because it is, as you point out, usually a stupid idea. I can see three >>reasons why it might be done: > >> - Someone is stupid in general. >> - Someone is not stupid in general, but making a stupid mistake. >> - Someone wants precisely that: shadow a set of one files with a set >> of different files. This might be useful in some situations. > >That third reason is important. An example: > >On HP-UX systems, there is a system management tool called SAM, the binaries >for which live in a directory on /usr. Now, on any decent system, /usr is >going to live on a different filesystem than /, so in single-user mode it may >not be available--like when moving /usr to a larger filesystem. Except with HP-UX 9 on a workstation which didn't support partitioning of disks on 700 series machines (ugh). >So I, at least, put a copy of /usr/sam (and a couple of other tools) in the >/usr directory which is the mountpoint for the /usr filesystem. Made my life >a lot simpler, and I hardly think it was stupid--it became SOP for a large >(then $750,000,000) networking company I worked for at the time... >-- >Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 > Current maintainer, MIT TECO EMACS (v. 170) >last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now Rich -- It seems reasonable to me... Must be that commercial experience. Comes from dealing with real world disasters on sites that do real work (not small college or development labs...) You see, when I taught Sysadmin for a major unix vendor I discussed stashing copies of most of the Unix text based databases (printcap, passwd, shadow, fstab under usr) in case something got seriously splattered. This would make recovery from Sysadmin error easier (they'd be on the root level 0 dump tape as well as instantly available in the event of sysadmin problems with an umount of /usr)... Bill ex-DEC Field Servant now doing Sysadmin for a living. This was done by a number of people at large commercial sites. ###### Message-ID: <367549EE.13B7@gazonk.del> Date: Mon, 14 Dec 1998 12:25:02 -0500 From: "Foobar T. Clown" Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <74mmn9$4l7$5@irk.zetnet.co.uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 14 Dec 1998 12:26:33 -0500, 198.133.110.49 Lines: 17 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!ix.netcom.com!zeus.nomos.com!198.133.110.49 David Wragg wrote: > > Libc is part of the OS by any reasonable standard. The typical libc is a big bag containg routines from several unrelated categories: - ANSI standard library routines (e.g., printf(), scanf(), strcpy()) These are part of the C programming language standard. - Run-time support for the compiler. (e.g., multiple precision arithmetic, initialization). These are part of the compiler. - non-standard, platform specific routines, - System calls (e.g., read(), write(), ioctl()) These, and only these, belong to the operating system. ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 14 Dec 1998 16:49:36 -0000 Lines: 24 Message-ID: <367541b4.0@192.168.0.20> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> NNTP-Posting-Host: 195.147.134.182 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 Hugh Davies wrote in message <74qpdt$57u@axalotl.demon.co.uk>... >In article >hshubs@mindspring.com (Howard S Shubs) writes: > >>Granted. Still, when I had to get an economist and an accountant used >>to UNIX (which isn't yet a completed project), I had -lots- of fun >>explaining certain things, like the chmod protection mask. > >So don't do it that way. What's wrong with > >chmod u+x,go-rwx filename I prefer right-click menu properties option choose what flags you want to set on/off Samael ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 14 Dec 1998 17:42:50 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 18 Message-ID: <753imq$1jg$1@agate.berkeley.edu> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <367541b4.0@192.168.0.20> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <367541b4.0@192.168.0.20>, Samael wrote: >>chmod u+x,go-rwx filename > >I prefer > >right-click menu >properties option >choose what flags you want to set on/off Try changing the permission of every file on your system that starts with 'a' and ends with 'e'. Or all the ".exe" files in a directory for that matter. See you tomorrow. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: "Bradley S. Bishop" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 14 Dec 1998 19:31:06 GMT Organization: Lucent Technologies, Columbus, Ohio Lines: 16 Message-ID: <01be2798$4885e7c0$159876c7@ga0235bbishop> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> NNTP-Posting-Host: 199.118.152.* X-Newsreader: Microsoft Internet News 4.70.1162 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!wn3feed!135.173.83.25!wn4feed!worldnet.att.net!207.24.196.41!nntphub.cb.lucent.com!news Actually, under Windows 95/NT this would be quite easy (I assume this is a debate on Unix vs. Windows now - came in on the tail end of it). Do a find (Start->Find->Files/Folders), select whichever directory you want to start in, type in a*e (or a*.*e -depends on what you are looking for). Select all (CTRL-A) and right-click->properties. Now change whatever attributes you wish. Eric J. Korpela wrote in article <753imq$1jg$1@agate.berkeley.edu>... > > Try changing the permission of every file on your system that starts with > 'a' and ends with 'e'. Or all the ".exe" files in a directory for that > matter. See you tomorrow. Brad ###### From: mvredenburg Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> Organization: Global 2000 Communications, Inc. Lines: 24 X-Newsreader: TIN [UNIX 1.3 unoff BETA release 961005] Message-ID: NNTP-Posting-Host: 205.247.154.34 X-Trace: news14.ispnews.com 913667657 205.247.154.34 (Mon, 14 Dec 1998 15:34:17 EDT) NNTP-Posting-Date: Mon, 14 Dec 1998 15:34:17 EDT Date: Mon, 14 Dec 1998 20:34:17 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!209.44.33.119!hub1.ispnews.com!news14.ispnews.com.POSTED!not-for-mail In alt.folklore.computers Bradley S. Bishop wrote: > Actually, under Windows 95/NT this would be quite easy (I assume this is a > debate on Unix vs. Windows now - came in on the tail end of it). Do a find > (Start->Find->Files/Folders), select whichever directory you want to start > in, type in a*e (or a*.*e -depends on what you are looking for). Select all > (CTRL-A) and right-click->properties. Now change whatever attributes you > wish. > Eric J. Korpela wrote in article > <753imq$1jg$1@agate.berkeley.edu>... > > > > Try changing the permission of every file on your system that starts with > > 'a' and ends with 'e'. Or all the ".exe" files in a directory for that > > matter. See you tomorrow. > Brad Will Win-95/NT expand a*e to a??????e or a??????? DOS 6.3 expands a*e to a??????? -- Myles Vredenburg mvredenburg at global2000 dot net ###### From: korpela@islay.ssl.berkeley.edu (Eric J. Korpela) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 14 Dec 1998 20:53:34 GMT Organization: Cal Berkeley-- Space Sciences Lab Lines: 34 Message-ID: <753tse$91s$1@agate.berkeley.edu> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> NNTP-Posting-Host: islay.ssl.berkeley.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!agate!islay.ssl.berkeley.edu!korpela In article <01be2798$4885e7c0$159876c7@ga0235bbishop>, Bradley S. Bishop wrote: >Eric J. Korpela wrote: >> Try changing the permission of every file on your system that starts with >> 'a' and ends with 'e'. Or all the ".exe" files in a directory for that >> matter. See you tomorrow. > >Actually, under Windows 95/NT this would be quite easy (I assume this is a >debate on Unix vs. Windows now - came in on the tail end of it). Do a find >(Start->Find->Files/Folders), select whichever directory you want to start >in, type in a*e (or a*.*e -depends on what you are looking for). Select all >(CTRL-A) and right-click->properties. Now change whatever attributes you >wish. I'm not sure I would call that easy. Definitely not quick given how slow the windows find is. For more of a challenge, do it automatically every night at 1am. It's actually a UNIX versus VMS debate, but someone with a windows attitude wanted to indicate how much "easier" windows is than unix "chmod" or VMS "set file /prot". Frankly either of them are better than navigating 3 menu levels with the mouse, switching to the keyboard and typing in the search pattern, waiting 5 minutes for windows to go through the 18000 files on a typical hard drive, hitting control-a and switching back to the mouse selecting the right tab on the properties menu and clicking up to 9 times to set the desired permissions. Give me 'find / -name "a*e" -exec chmod go-rwx {} \; &' any day. Eric -- Eric Korpela | An object at rest can never be korpela@ssl.berkeley.edu | stopped. Click for home page. ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 14 Dec 1998 22:31:33 -0600 Organization: MindSpring Enterprises Lines: 13 Message-ID: References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <7523bh$cm$1@daemonweed.reanimators.org> NNTP-Posting-Host: d1.56.d1.7e X-Server-Date: 15 Dec 1998 04:31:34 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <7523bh$cm$1@daemonweed.reanimators.org>, Frank McConnell wrote: >It confuses folks like me who have had enough exposure to VMS that we >have to sit there and think about whether "o" means "owner" like it >does in SET FILE/PROT. Generally speaking, by the time I remember I >worked out what I want in octal anyway. Exactly. Since the people I'm training are coming from an MVS background, they won't have the same problems. I'm planning on trying his concept soon. -- Howard S Shubs The Denim Adept ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: lucvdv@null.net's message of Sat, 12 Dec 1998 16:28:55 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> <36728afd.10991534@news.uunet.be> Date: Tue, 15 Dec 1998 00:19:50 GMT Lines: 23 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!194.165.93.117.MISMATCH!newshub.bart.net!newsfeed.wirehub.nl!newspeer1.nac.net!netnews.com!ix.netcom.com!netcom!alderson In article <36728afd.10991534@news.uunet.be> lucvdv@null.net (Luc Van der Veken) writes: >Also sprach alderson@netcom.netcom.com (Richard M. Alderson III) >on Fri, 11 Dec 1998 20:03:04 GMT to alt.folklore.computers: >>Unless Linux is very different from the Unices I've worked with (SunOS, >>HP-UX, Solaris, Ultrix, OSF/1, Digital Unix 4.0, Pyramid), you can't copy >>files to an unmounted filesystem, so this can't work. >You can temporarily mount it on another mount point just for copying. How else >would you - when setting it up - get all the existing files in there in the >first place? That's an awful lot of trouble to go to just to save a minuscule amount of disk space, especially since the OS isn't going to let you umount /usr and /etc while you're running. The only way what was originally suggested makes any sense at all is if you *could* write to unmounted filesystems. Anything else is just schratzing around for the sake of schratzing around. -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 00:58:43 +0100 Organization: [Posted via] Interactive Networx Lines: 16 Message-ID: <1dk1aof.15bij1l14q95uiN@n31-109.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <74knnt$8c1@news1.newsguy.com> <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <74tckl$al@axalotl.demon.co.uk> <74v2ap$pbc$1@news1.bu.edu> NNTP-Posting-Host: n31-109.berlin.snafu.de X-Newsreader: MacSOUP 2.3.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!194.162.162.196!newsfeed.nacamar.de!news-hh.maz.net!unlisys!news.snafu.de!jnickelsen Howard S Shubs wrote: [Unix permission bits and symbolc representation] > I find the qualifiers for u-ser, g-roup, and o-ther to be kinda tough [...] *Years* I confusedly took "o" for "owner", getting annoyed and having to correct my mistake each time . Things got better when, for some time, I often had to make files go-w. I made an alias, shell function, or script (whatever, I don't remember) to set these permissions, named, of course, "go-w". And, see, here we have a fine example how Unix commands got their legendary intuitive and consistent names. ;-) -- Juergen Nickelsen ###### From: mww@microfocus.com (Michael Wojcik) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 15 Dec 1998 02:14:25 GMT Organization: Micro Focus Inc. Lines: 86 Message-ID: <754gm1$b1@news1.newsguy.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <366b39eb.3006571501@news.cyberport.com> <74glh7$j1n$1@strato.ultra.net> <74kkr4$6t9@news1.newsguy.com> <74oi4r$nrp$4@ligarius.ultra.net> <74p9fu$a0n@news3.newsguy.com> <74u04s$4kf$3@ligarius.ultra.net> NNTP-Posting-Host: p-682.newsdawg.com X-Newsreader: xrn 9.00 Originator: mww@raederle.microfocus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!pln-e!spln!extra.newsguy.com!newsp.newsguy.com!mww In article <74u04s$4kf$3@ligarius.ultra.net>, jmfbahciv@aol.com writes: > In article <74p9fu$a0n@news3.newsguy.com>, > mww@microfocus.com (Michael Wojcik) wrote: > > > >In article <74oi4r$nrp$4@ligarius.ultra.net>, jmfbahciv@aol.com writes: > >> In article <74kkr4$6t9@news1.newsguy.com>, > >> mww@microfocus.com (Michael Wojcik) wrote: > >Note that interrupt handlers, while preemptible (by kernel processes > >at higher priority), are not pageable and presumably not reentrant. > > But interrupts are done on a per CPU basis. So that doesn't make > any sense. There definitely will be code that can't be interrupted > until the data necessary to handle the interrupt is stored in a safe > place for the code that will take care of the event that caused the > interrupt. But that's a very, very small code window (or should be > for performance's sake, anyway). Right. Interrupt handlers in AIX are required (by design guidelines; I don't know if there are mechanisms in the kernel to enforce this) to be minimal pieces of code that just shift data across the bus or whatever. Most of a device driver is pageable and reentrant. > >Performance is fine, unless you run out of virtual memory; but > >thrashing is thrashing, and AIX does have a mechanism for recovering > >more or less gracefully from memory exhaustion (it starts killing > >off processes in some "least-favored" order, first with a catchable > >signal so the process can exit cleanly and then with an uncatchable > >one to enforce it, until it's below the danger mark). > > One of the things that I've been forgetting is that hardware > nowadays isn't as slow as in my day. So the bottlenecks move > to the slowest as hardware is upgraded. This was also true > when TW would do the device drivers for new hardware. Putting > a faster disk on a system abruptly caused the same user mix > to become CPU bound. Putting on a faster CPU would invariably > cause the same mix to become I/O bound. [...] Sure. Also, note that the VMM code isn't even executed "most" of the time, unless the machine is seriously overcommitted and thrashing; it only gets invoked to handle paging. With RAM as cheap as it is these days, typical Unix systems page relatively infrequently. Right now my machine is using nearly all 64 MB of RAM plus 73 MB of disk swap space, but it's averaging about 6 page faults a *minute*. The VMM is doing a good job of keeping the most frequently used pages in RAM. > How many jobs typically run on one of those machines? They can have anywhere from a handful of active processes to thousands. Unix machines run from (multitasking) laptops to some of the largest computers in the world, running big parallel number-crunching jobs. Many have hundreds of users; I've never actually seen a machine with thousands of signed-on users, but I've heard customers suggest they might be scaling up to that level, so someone's probably there already. IBM has various web pages bragging about this and that big installation, like the other players in the market. But of course these days it's customary to build machines with multiple gigabytes of RAM. You can fit a lot of user jobs in that kind of space. > >The problem with Unix is that the panic may have been issued by > >third-party device driver code or kernel extensions. Who's to say > >what's been damaged? > > The damage detector. Yes, there's a lot of room for improvement in kernel self-monitoring in all the Unix implementations I've seen. But it's been years since I looked at a Unix kernel in depth, and it's possible that designs are getting more robust. Michael Wojcik mww@microfocus.com AAI Development, Micro Focus Inc. Department of English, Miami University Q: What is the derivation and meaning of the name Erwin? A: It is English from the Anglo-Saxon and means Tariff Act of 1909. -- Columbus (Ohio) Citizen ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <366B7CFA.2A976248@tnglwood.demon.co.uk> <36728afd.10991534@news.uunet.be> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 20 Message-ID: Date: Tue, 15 Dec 1998 04:58:47 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913697927 205.166.146.8 (Mon, 14 Dec 1998 22:58:47 CDT) NNTP-Posting-Date: Mon, 14 Dec 1998 22:58:47 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article , Richard M. Alderson III wrote: >That's an awful lot of trouble to go to just to save a minuscule amount of disk >space, especially since the OS isn't going to let you umount /usr and /etc >while you're running. The only way what was originally suggested makes any >sense at all is if you *could* write to unmounted filesystems. Anything else >is just schratzing around for the sake of schratzing around. I would do it the other way; do one of the loopback mount type of things (doesn't every Unix have about three of these nowadays?) and mount the "underlying" /etc as, say, '/var/boot/etc', and then mount a "real" /etc over it. If the loopback mount does the "right" thing, you'll have the "underlying" contents and the "real" contents at the same time. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <751rra$11i4@i4got.pechter.dyn.ml.org> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 21 Message-ID: Date: Tue, 15 Dec 1998 05:00:04 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913698004 205.166.146.8 (Mon, 14 Dec 1998 23:00:04 CDT) NNTP-Posting-Date: Mon, 14 Dec 1998 23:00:04 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <751rra$11i4@i4got.pechter.dyn.ml.org>, Bill Pechter wrote: >You see, when I taught Sysadmin for a major unix vendor I discussed >stashing copies of most of the Unix text based databases >(printcap, passwd, shadow, fstab under usr) in case something got >seriously splattered. This would make recovery from Sysadmin error >easier (they'd be on the root level 0 dump tape as well as instantly >available in the event of sysadmin problems with an umount of /usr)... When I worked for a not-so-major, but very excellent, unix vendor, they automatically did this, and called it something like "/var/backups". Generally, '/var' is either its own mount point, or '/usr/var', so this works. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <74qq3m$57u@axalotl.demon.co.uk> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 15 Message-ID: Date: Tue, 15 Dec 1998 05:08:06 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913698486 205.166.146.8 (Mon, 14 Dec 1998 23:08:06 CDT) NNTP-Posting-Date: Mon, 14 Dec 1998 23:08:06 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <74qq3m$57u@axalotl.demon.co.uk>, Hugh Davies wrote: >Hmmm. I have "Another Evening Wasted with Tom Lehrer", as well. Could you post reference info so I can order it, if I can ever find a company to order from that doesn't spam? ;-) (Actually, I believe CDNow is okay.) -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 06:57:01 +0000 Organization: Tanglewood Message-ID: <3676083D.E3938AB2@tnglwood.demon.co.uk> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913706349 nnrp-11:9188 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 17 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!195.252.142.12!newsfeed.tli.de!newsfeed.wirehub.nl!bullseye.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Eric J. Korpela wrote: > I'm not sure I would call that easy. Definitely not quick given how slow > the windows find is. For more of a challenge, do it automatically every > night at 1am. Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, Windows is the other way around. I suspect that there would be a case for a graphical front end to a system with proper scripting, to help those of us who are getting older, and can't always remember all the options. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### Date: 15 Dec 98 10:17:45 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> Message-ID: <1704.653T2288T6175773@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 37 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 15 Dec 1998 10:18:41 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.127 In article <755h9n$deh$1@ligarius.ultra.net> jmfbahciv@aol.com (jmfbahciv) writes: >In article <3676083D.E3938AB2@tnglwood.demon.co.uk>, >Robert Billing wrote: > >>Eric J. Korpela wrote: >> >>> I'm not sure I would call that easy. Definitely not quick >>>given how slow the windows find is. For more of a challenge, >>>do it automatically every night at 1am. >> >> Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually >>easy, Windows is the other way around. I suspect that there would be >>a case for a graphical front end to a system with proper scripting, >>to help those of us who are getting older, and can't always remember >>all the options. > >Right. It's been a puzzle to me why people think these menu thingies >are better. Even JMF (who typed with his two index fingers and one >thumb) could do the data manipulations with less key strokes than >those required when using a mouse. I would much, much rather >type than click--typing is infinitely "more user-friendly". The paradox arises because most people don't realize that "easy to learn" is not the same as "easy to use". Menus are great if you're noodling around trying to get the lay of the land, but they just get in the way once you know what you're doing. Of course, nowadays your average luser never gets to know what he's doing, either because he doesn't use a program frequently enough to learn it, or because he refuses to learn, or because he quickly throws the program away in favour of next month's fad. Thus the myth that command lines are "difficult to use" gets perpetuated. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Date: 15 Dec 98 10:17:45 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> Message-ID: <1704.653T2288T6175773@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 37 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 15 Dec 1998 10:18:41 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.127 In article <755h9n$deh$1@ligarius.ultra.net> jmfbahciv@aol.com (jmfbahciv) writes: >In article <3676083D.E3938AB2@tnglwood.demon.co.uk>, >Robert Billing wrote: > >>Eric J. Korpela wrote: >> >>> I'm not sure I would call that easy. Definitely not quick >>>given how slow the windows find is. For more of a challenge, >>>do it automatically every night at 1am. >> >> Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually >>easy, Windows is the other way around. I suspect that there would be >>a case for a graphical front end to a system with proper scripting, >>to help those of us who are getting older, and can't always remember >>all the options. > >Right. It's been a puzzle to me why people think these menu thingies >are better. Even JMF (who typed with his two index fingers and one >thumb) could do the data manipulations with less key strokes than >those required when using a mouse. I would much, much rather >type than click--typing is infinitely "more user-friendly". The paradox arises because most people don't realize that "easy to learn" is not the same as "easy to use". Menus are great if you're noodling around trying to get the lay of the land, but they just get in the way once you know what you're doing. Of course, nowadays your average luser never gets to know what he's doing, either because he doesn't use a program frequently enough to learn it, or because he refuses to learn, or because he quickly throws the program away in favour of next month's fad. Thus the myth that command lines are "difficult to use" gets perpetuated. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 11:37:16 -0000 Lines: 57 Message-ID: <36764a3a.0@192.168.0.20> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> NNTP-Posting-Host: 195.147.134.193 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 Eric J. Korpela wrote in message <753tse$91s$1@agate.berkeley.edu>... >In article <01be2798$4885e7c0$159876c7@ga0235bbishop>, >Bradley S. Bishop wrote: >>Eric J. Korpela wrote: >>> Try changing the permission of every file on your system that starts with >>> 'a' and ends with 'e'. Or all the ".exe" files in a directory for that >>> matter. See you tomorrow. >> >>Actually, under Windows 95/NT this would be quite easy (I assume this is a >>debate on Unix vs. Windows now - came in on the tail end of it). Do a find >>(Start->Find->Files/Folders), select whichever directory you want to start >>in, type in a*e (or a*.*e -depends on what you are looking for). Select all >>(CTRL-A) and right-click->properties. Now change whatever attributes you >>wish. > >I'm not sure I would call that easy. Definitely not quick given how slow >the windows find is. For more of a challenge, do it automatically every >night at 1am. Scripting is a tad harder (obviously), but perfectly possible with the windows scripting tool. > >"set file /prot". Frankly either of them are better than navigating 3 >menu levels with the mouse, switching to the keyboard and typing in the >search pattern, F3 brings up the search window. waiting 5 minutes for windows to go through the 18000 files >on a typical hard drive, 8 seconds to search my 3 Gig hard drive (nearly full - contains about 20000 files) >hitting control-a and switching back to the mouse >selecting the right tab on the properties menu and clicking up to 9 times to >set the desired permissions. How about Ctrl-A, the windows key (to bring up the properties window) and then using the mouse to set up whatever properties you want to). > >Give me 'find / -name "a*e" -exec chmod go-rwx {} \; &' any day. Fair enough, once you understand how pipes work, passing parameters between programs and the different properties that the different programs take. New Users (and most old users) don't need that kind of trouble. Samael ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 98 13:03:27 GMT Organization: UltraNet Communications, Inc. Lines: 39 Message-ID: <755o3t$66l$2@strato.ultra.net> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <36764a3a.0@192.168.0.20> NNTP-Posting-Host: d9.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 15 Dec 1998 13:27:25 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!feed1.news.rcn.net!rcn!news.ultranet.com!d9 In article <36764a3a.0@192.168.0.20>, "Samael" wrote: > >Eric J. Korpela wrote in message <753tse$91s$1@agate.berkeley.edu>... >>In article <01be2798$4885e7c0$159876c7@ga0235bbishop>, >>Bradley S. Bishop wrote: >>>Eric J. Korpela wrote: >>>> Try changing the permission of every file on your system that starts >with >>>> 'a' and ends with 'e'. Or all the ".exe" files in a directory for that >>>> matter. See you tomorrow. >>> >>>Actually, under Windows 95/NT this would be quite easy (I assume this is a >>>debate on Unix vs. Windows now - came in on the tail end of it). Do a find >>>(Start->Find->Files/Folders), select whichever directory you want to start >>>in, type in a*e (or a*.*e -depends on what you are looking for). Select >all >>>(CTRL-A) and right-click->properties. Now change whatever attributes you >>>wish. >> >>I'm not sure I would call that easy. Definitely not quick given how slow >>the windows find is. For more of a challenge, do it automatically every >>night at 1am. > >Scripting is a tad harder (obviously), but perfectly possible with the >windows scripting tool. Sigh! But clicking can't be automated with a batch job slated to run every night when all the critters are in bed. /BAH Subtract a hundred and four for e-mail. ###### From: J.Hendrickx@maw.kun.nl (J.Hendrickx) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 14:49:03 +0100 Organization: KUN Lines: 20 Message-ID: References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> NNTP-Posting-Host: mwpc87.daf.kun.nl X-Newsreader: MicroPlanet Gravity v2.10 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newspump.monmouth.com!newspeer.monmouth.com!cyclone.news.idirect.com!island.idirect.com!newsfeed.wirehub.nl!surfnet.nl!barba.uci.kun.nl!not-for-mail In article <3676083D.E3938AB2@tnglwood.demon.co.uk>, unclebob@tnglwood.demon.co.uk says... > Eric J. Korpela wrote: > > > I'm not sure I would call that easy. Definitely not quick given how slow > > the windows find is. For more of a challenge, do it automatically every > > night at 1am. > > Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, > Windows is the other way around. I suspect that there would be a case > for a graphical front end to a system with proper scripting, to help > those of us who are getting older, and can't always remember all the > options. > For one or two files, Windows is easy. For several files, the old DOS attrib would be better. So you can have it both ways. A macro program can schedule events for certain times, at least mine can. Ease of use isn't Windows weak spot. Ask me how I can be sure Windows will be up and running at 1am (or some other point in time when I'm not around to reboot it). ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 15 Dec 1998 15:13:58 GMT Message-ID: <755ubm$quj$1@irk.zetnet.co.uk> References: NNTP-Posting-Host: man-047.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 913734838 27603 194.247.41.58 (15 Dec 1998 15:13:58 GMT) NNTP-Posting-Date: 15 Dec 1998 15:13:58 GMT X-Everything: Net-Tamer V 1.08X Lines: 29 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!nntp.abs.net!newsfeed.ecrc.net!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-09 meowing@banet.net(Fluffy) said: :> >Hmm. If Unix style shell globbing is a Good Idea, why do I so :>often >find myself having to rely on ugly hacks like xargs to run :>things >against large sets of files? :> Because it wasn't done quite right. It's still a Good Idea, :though. :But when that filespac is just used a a param called by a function :at runtime, the problem is solved. And that's what VMS lets you do. :> Don't confuse a concept with its implementation. :I haven't. Hell, they both have good points and bad points, don't they? In Unix, no program ever has to worry about expanding filenames, because it's done for them; all the program has to do is go down its list of files and do things with each. The globbing might not always do what you want, but most of the time it works pretty well. Whereas the other alternative just passes in the mask to the application and leaves it to handle the globbing. Great, it lets them find associated files; but ill-behaved program will refuse to glob at all, and the only way round that is to have the shell do it. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 15:19:16 -0000 Lines: 37 Message-ID: <36767e0a.0@192.168.0.20> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> NNTP-Posting-Host: 195.147.133.25 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 jmfbahciv@aol.com wrote in message <755h9n$deh$1@ligarius.ultra.net>... >In article <3676083D.E3938AB2@tnglwood.demon.co.uk>, > Robert Billing wrote: >>Eric J. Korpela wrote: >> >>> I'm not sure I would call that easy. Definitely not quick >>>given how slow the windows find is. For more of a challenge, >>>do it automatically every night at 1am. >> >> Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, >>Windows is the other way around. I suspect that there would be a case >>for a graphical front end to a system with proper scripting, to help >>those of us who are getting older, and can't always remember all the >>options. >> >Right. It's been a puzzle to me why people think these menu thingies >are better. Even JMF (who typed with his two index fingers and one >thumb) could do the data manipulations with less key strokes than >those required when using a mouse. I would much, much rather >type than click--typing is infinitely "more user-friendly". It's not a speed thing, it's a memory thing. If I want to move some text from one place t another, it's erasy to look at the menus, spot '"Edit' and then look down the listun til I see the word 'Cut' and think, aah, that's what I want, than try and remember any key combinations. Same with a file copy. Select a file, right click and look down the list of files until I see the option I want (copy). Then go to where I want it and say 'Paste'. No memory needed. Still slightly unintuitive (until the user gets used to everything in windows being cut and paste to move it around), but a lot easier than remembering directory names and where everything is. Samael ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 15:21:33 -0000 Lines: 50 Message-ID: <36767e92.0@192.168.0.20> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <36764a3a.0@192.168.0.20> <755o3t$66l$2@strato.ultra.net> NNTP-Posting-Host: 195.147.133.25 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 jmfbahciv@aol.com wrote in message <755o3t$66l$2@strato.ultra.net>... >In article <36764a3a.0@192.168.0.20>, > "Samael" wrote: >> >>Eric J. Korpela wrote in message <753tse$91s$1@agate.berkeley.edu>... >>>In article <01be2798$4885e7c0$159876c7@ga0235bbishop>, >>>Bradley S. Bishop wrote: >>>>Eric J. Korpela wrote: >>>>> Try changing the permission of every file on your system that starts >>with >>>>> 'a' and ends with 'e'. Or all the ".exe" files in a directory for >that >>>>> matter. See you tomorrow. >>>> >>>>Actually, under Windows 95/NT this would be quite easy (I assume this is >a >>>>debate on Unix vs. Windows now - came in on the tail end of it). Do a >find >>>>(Start->Find->Files/Folders), select whichever directory you want to >start >>>>in, type in a*e (or a*.*e -depends on what you are looking for). Select >>all >>>>(CTRL-A) and right-click->properties. Now change whatever attributes you >>>>wish. >>> >>>I'm not sure I would call that easy. Definitely not quick given how >slow >>>the windows find is. For more of a challenge, do it automatically every >>>night at 1am. >> >>Scripting is a tad harder (obviously), but perfectly possible with the >>windows scripting tool. > > >Sigh! But clicking can't be automated with a batch job slated >to run every night when all the critters are in bed. That's why you'd use the scripting tool for that kind of thing. It doesn't involve clicking. Alternatively, I know a couple of sysadmins on NT servers that knock together a simple VB program to copy files back and forh when necessary. Tad kludgy, but it's a reasonable CLI make-do. Alternatively, you can use the Command Line in NT, which emulates DOS and use Batch files for your processing and automate them with the Windwos Task Scheduler. Samael ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 15 Dec 1998 16:06:51 GMT Organization: University of Illinois at Urbana-Champaign Lines: 10 Message-ID: <7561er$o2b$1@vixen.cso.uiuc.edu> References: <1456.648T1620T6105137@sky.bus.com> <753tse$91s$1@agate.berkeley.edu> <36764a3a.0@192.168.0.20> <755o3t$66l$2@strato.ultra.net> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!128.174.5.49!vixen.cso.uiuc.edu!tsmurphy In article <755o3t$66l$2@strato.ultra.net>, wrote: >Sigh! But clicking can't be automated with a batch job slated >to run every night when all the critters are in bed. Of course it can. Just use one of those GUI recorders, record what you need, and put it in a scheduler to run every night. I don't see why this is so difficult. -- Terry ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 17:13:38 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 26 Message-ID: <36769835.1945128@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <748j4a$lfi$2@strato.ultra.net> <36691be2.2492303@news.uunet.be> <366b3255.941270842@Rockyd> <749co3$j9h@srvr1.engin.umich.edu> <36683c0a.1813126@Rockyd> <366858fd.4481731@news.vip.net> <366913D1.986841E5@tnglwood.demon.co.uk> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!newsfeed.nacamar.de!news-hh.maz.net!unlisys!news.snafu.de!home.ibert.com!news On Sat, 05 Dec 1998 11:06:57 +0000, Robert Billing wrote: > Yes, and how it does it is very clever. Basically (scratches head, >tries to remember DM from BTDOT) VMS knows when a page has been written >to by a process, and so knows that some pages don't match what is on >disk, these pages have to be rewritten if they are to go out of memory. >What it then does is remove pages from process working sets, but leave >them in memory. If the process asks for them again, it simply reattaches >them to the working set, if the process doesn't ask for them back, then >in time they are treated as free memory and can be allocated to other >processes. If the original process asks for the page again, then new >memory is allocated, and the page read in again from disk. What this >means in practice is that pages which aren't being used (such as >initialisation code) get nibbled out of the working set, leaving only >the active pages in memory. Um ... that is basically a quite decent descriptions of how modern UN*Xoid operating system manage memory pages, right? At least I remember having read a description in at least one of the Garden Book, the old and the new Dragon Book, and the paper on the new VM system for BSD. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### Newsgroups: alt.folklore.computers From: jwbirdsa@picarefy.picarefy.com (James W. Birdsall) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec15.190813.10619@picarefy.picarefy.com> Organization: Green Tiger Software References: <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <36767e0a.0@192.168.0.20> Date: Tue, 15 Dec 1998 19:08:13 GMT Lines: 20 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.cwix.com!199.181.164.14!news.seanet.com!nntp.picarefy.com!picarefy!jwbirdsa In article <36767e0a.0@192.168.0.20> "Samael" writes: >It's not a speed thing, it's a memory thing. If I want to move some text >from one place t another, it's erasy to look at the menus, spot '"Edit' and >then look down the listun til I see the word 'Cut' and think, aah, that's >what I want, than try and remember any key combinations. If you have to remember key combinations, you haven't learned the program yet. When you have, as soon as you know what you want to do, your hands will do it, in the same way that you type text without thinking about hitting the individual keys. This is why so many people want keyboard shortcuts. Menus *are* handy for infrequently-used functions, but being forced to go through them just slows down experienced users. -- James W. Birdsall http://www.picarefy.com/~jwbirdsa/ jwbirdsa@picarefy.com "For it is the doom of men that they forget." -- Merlin Get the Sun-2 Hardware Reference from ftp.picarefy.com:/pub/Sun-Hardware-Ref Sun-2 Hardware Reference Web Page: http://sun-www.picarefy.com/ ###### From: Mark Wilkins Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 15:31:23 -0800 Organization: Walt Disney Feature Animation Lines: 43 Message-ID: <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> NNTP-Posting-Host: 172.30.23.178 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.06 [en] (X11; I; IRIX 6.2 IP22) To: Charlie Gibbs Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news-peer.gip.net!news.gsl.net!gip.net!news.idt.net!nntp2.cerf.net!nntp3.cerf.net!louie.disney.com!not-for-mail Charlie Gibbs wrote: > The paradox arises because most people don't realize that "easy to > learn" is not the same as "easy to use". Menus are great if you're > noodling around trying to get the lay of the land, but they just get > in the way once you know what you're doing. Here's a different perspective: I work with a number of people who have essentially an artistic background and have learned to use computers. I also work with a number of people from a software engineering background. A key difference between these two groups of people is that the artists have a much stronger memory for spatial relationships and a much weaker memory for verbal syntaxes. Remembering which of two commands to use in a command-line environment is a verbal game, essentially, while remembering which command serves a given function in a graphic user interface is a matter of spatial memory. My point is that the artists often find graphic menus far easier to use even after they've achieved a high level of proficiency with a software package because the means of interaction is suited to their strengths. Meanwhile, software engineers find themselves more comfortable with command lines and often ask why graphic interfaces are even necessary. Many of the artists require repeated explanations to remember a textual command but only need to be shown a menu item once. What is a fallacy is to suggest that to have difficulty with textual commands is equivalent to failure to learn the software or achieve the state of "knowing what you're doing." Though these artists may not have the same strengths in symbolic manipulation that I do, they are able to use their spatial memory to be productive in ways I could never dream of. -- Mark P.S. The Walt Disney company doesn't have anything to say about the above issues. Just me. :-) ----------------------------------------------------------- Mark Wilkins Walt Disney Feature Animation mark_wilkins @fa.disney.com Burbank, CA ###### Date: 15 Dec 98 17:31:54 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> Message-ID: <2230.653T2739T10516280@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 47 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 15 Dec 1998 17:51:02 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-pen-3.sprintlink.net!news-east1.sprintlink.net!news-peer1.sprintlink.net!news-backup-west.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.115 In article <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> Look@the_bottom_for_my.e-mail (Mark Wilkins) writes: >Charlie Gibbs wrote: > >> The paradox arises because most people don't realize that "easy to >> learn" is not the same as "easy to use". Menus are great if you're >> noodling around trying to get the lay of the land, but they just get >> in the way once you know what you're doing. > > Here's a different perspective: > > I work with a number of people who have essentially an artistic >background and have learned to use computers. I also work with a >number of people from a software engineering background. > > A key difference between these two groups of people is that the >artists have a much stronger memory for spatial relationships and >a much weaker memory for verbal syntaxes. Remembering which of >two commands to use in a command-line environment is a verbal game, >essentially, while remembering which command serves a given function >in a graphic user interface is a matter of spatial memory. > > My point is that the artists often find graphic menus far easier >to use even after they've achieved a high level of proficiency with >a software package because the means of interaction is suited to >their strengths. Meanwhile, software engineers find themselves more >comfortable with command lines and often ask why graphic interfaces >are even necessary. Many of the artists require repeated explanations >to remember a textual command but only need to be shown a menu item >once. > > What is a fallacy is to suggest that to have difficulty with textual >commands is equivalent to failure to learn the software or achieve the >state of "knowing what you're doing." Though these artists may not >have the same strengths in symbolic manipulation that I do, they are >able to use their spatial memory to be productive in ways I could >never dream of. A fascinating alternate view. Thanks for pointing that out. You've illustrated two totally different forms of abstraction. And I'd say that's one of the most convincing arguments yet that a system should provide both textual and graphic interfaces. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <755ubm$quj$1@irk.zetnet.co.uk> From: Marco S Hyman Date: 15 Dec 1998 17:52:43 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 Lines: 26 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 913773164 218 marc@204.94.187.130 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news2.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail alderson@netcom.netcom.com (Richard M. Alderson III) writes: > Your prejudices are showing. No one writing a file-handling program that > expects multi-file specifications in a system in which wildcard expansion > is handled by system calls would ignore those system calls. Snicker! As someone once said "never underestimate the power of human stupidity!" Just because a system call or library function exists does NOT mean it will be used. Or have you never seen this code (or something very like it) in random applications: for ( i = 0; i < count; i++ ) { dst[ i ] = src[ i ]; } Seems like either bcopy/memcpy would do the job (and perhaps more efficiently). > And a Unix program that only expects a single filename won't do well with > globbing, either, now will it? Eh? bad code can be written for any operating system. If a program expects one and only argument it should damn well check that that's what it got. What's UNIX got to do with it? // marc ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 19:57:14 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 76 Message-ID: <3677bac7.10795833@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On Mon, 07 Dec 1998 10:07:15 -0600, hshubs@mindspring.com (Howard S Shubs) wrote: >1) poor command-line design Agreed. >2) lack of developer support other than the "make" command But that is something and blows MMS out of the water without even trying. >3) lack of portability I don't think I can follow you there. I can port a program I have written for any UN*Xoid (especially if I am sort of experienced and write it in such a way to avoid very special features the UN*Xoid in question might have) to any other UN*Xoid with little effort. But where can I port a VMS program I have written? To OpenVMS? >(2) Let me ask you about the UNIX source-level debuggers. So, if you get >a crash, you can use the "core" file. Great. Indeed. [more rambling, much of what is quite true, deleted] What I really like is Purify. Okay, it costs money (rather a lot of it), but it is really worth it. It will find almost every dynamic-memory-related problem in almost no time at all for you. >(3) So, if I take a binary from UNIX box X to UNIX box Y, it'll just >-work-, right? Y'know, like I can take a VMS binary from a 20-year-old >VAX and run it on the newest VAX. Ditto with VMS on an Alpha. > >Okay, if you disallow that because they both come from the same vendor, >how about language compilers at least? If I build arbitrary program A on >an HP/UX box, will I be able to build it WITHOUT ANY CHANGES on AIX or AUX >or Digital UNIX? The changes will be absolutely marginal compared with the changes needed to port your VMS program to _anywhere_. So I really don't see your point at all. Porting from VMS to VMS or VMS to OpenVMS doesn't count. >UNIX... no wonder it's being supplanted by Windows NT. Big wonder, technically speaking. I do understand that you route for NT because it is Dave Cutler's child, but from painful experience I can tell you that NT is an absolute bitch to develop for. - Not a trace of consistency in the Win32 API. You have to look up the return codes for every fucking call. What is the return value for error? NULL? FALSE? INVALID_HANDLE_VALUE? Non-Zero? Christ, that's a lot easier even in OS/2. With UNIX, it is essentially a no-brainer, once you get the general hang of it. -1 is error if the value is an int, NULL if it is a pointer. Exceptions are lamentable, but fairly rare, aberrations [can't think of any right now]. - The most feauritis-infested, bulky, unstable, unpleasant, klugy development environment ever. You don't like UN*X source-code debuggers? Try Microsoft's. You'll love gdb after that. (At least Microsoft Visual Nightmare doesn't take all night to crash.) And the list goes on. Believe me, I've been there, done that, etc. I have even developed for VMS once, and I kinda hated that as well, because C was very poorly supported back then. (Is it now?) The compiler was non-standard. (errno that is not an lvalue. Brrr.) If you looked up a system call in the orange wall, there were code examples galore. BLISS, Assembler, COBOL, you name it, they had it. But I don't recall ever seeing a VAX C example. String handling is a mess. Plus, there wasn't a way to read as many bytes from a file as stat would tell you the file holds. You can open in one way and get more. You open it another way and get less. Messy. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:06:39 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 18 Message-ID: <3678c0d1.12342645@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On 9 Dec 1998 16:34:55 GMT, tsmurphy@cs.uiuc.edu (TERENCE MURPHY) wrote: >Please note that they say "requires Solaris" not requires "Solaris plus >/bin/sh and /bin/ls and /bin/cp", etc. Therefore one can conclude >that the term "Solaris" means not the kernel but the kernel as well >as included utilities. It's like this for any Unix. Solaris is not an operating system (at least it wasn't before Solaris 7). I think they call it an "operating environment" - it contains an operating system (SunOS 5 in the case of Solaris 2) plus extra bits and pieces. But I do admit that a shell is included in the SunOS bit. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:09:58 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 14 Message-ID: <3679c199.12542453@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <366f0671.114112447@news.bright.net> <74n69q$3n4$1@vixen.cso.uiuc.edu> <36708455.93118759@news.bright.net> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On Fri, 11 Dec 1998 02:51:00 GMT, deke.spamblock@generous.net wrote: >Can you name a Unix that does not contain cc? Lots! Unfortunately, "unbundling" the compiler has become rather commonplace these days. Sometimes there is some program called cc that does not do anything besides tell you that since you didn't buy the unbundled development environment, it cannot help you in any way. Ah, the joys of open-source UN*Xoid operating systems ... -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:12:06 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 13 Message-ID: <367ac266.12747713@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <366d1f6c.9761743@news.bright.net> <74jjnt$i01$1@vixen.cso.uiuc.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <74qq52$rm5$2@hades.csu.net> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On 11 Dec 1998 09:54:42 GMT, jkatz@ultrix6.cs.csubak.edu (who?) wrote: >Careful... >poking about DU 4.0, when, I stumbled upon this: >myhost:~$ diff /bin/sh /usr/bin/posix/sh >Binary files /bin/sh and /usr/bin/posix/sh differ Of course. /usr/bin/posix/sh is probably dynamically linked, but /bin/sh had rather not be. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:14:49 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 13 Message-ID: <367cc314.12921262@news.home.ibert.com> References: <74jjnt$i01$1@vixen.cso.uiuc.edu> <74mmn1$4l7$3@irk.zetnet.co.uk> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On 9 Dec 1998 20:31:29 GMT, lisard@zetnet.co.uk wrote: >(The Bourne shell does have some advantages; it's a fairly lightweight >scripting platform, and it keeps out of the way as much as possible. But >I still don't think it's actually meant to be *used* by anyone. ;> ) It is. For running scripts. :-) The C shell, on the other hand, is not. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:21:42 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 11 Message-ID: <367dc49a.13311869@news.home.ibert.com> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.maz.net!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On 11 Dec 1998 09:42:21 GMT, huge@nospam.demon.co.uk (Hugh Davies) wrote: >So don't do it that way. What's wrong with >chmod u+x,go-rwx filename It still looks like random line noise to the average non-technical person, and the techie starts doing octal in his head ... :-) -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:24:21 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 14 Message-ID: <367ec532.13463432@news.home.ibert.com> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <7523bh$cm$1@daemonweed.reanimators.org> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.pop.de!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On 13 Dec 1998 20:14:41 -0800, Frank McConnell wrote: >It confuses folks like me who have had enough exposure to VMS that we >have to sit there and think about whether "o" means "owner" like it >does in SET FILE/PROT. Generally speaking, by the time I remember I >worked out what I want in octal anyway. My VMS exposure has been rather limited (thank God), but I still think "o for owner, g for group and o fo... shit!" once in a while. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:37:57 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 13 Message-ID: <367fc85e.14275421@news.home.ibert.com> References: <74mmn9$4l7$5@irk.zetnet.co.uk> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news.tvd.be!newsfeed.wirehub.nl!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On 9 Dec 1998 20:31:37 GMT, lisard@zetnet.co.uk wrote: >On the other hand, since a >directory is merely a list of filenames and i-nodes, you can open a >directory as a file, which eases the pain somewhat. You can, but you rather shouldn't. Gone are the days when directory entries were all 2 bytes i-node, 14 bytes name - you never know what you will find! -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:43:02 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 11 Message-ID: <3680c9be.14628000@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news.tvd.be!newsfeed.wirehub.nl!news-fra.maz.net!newscore.ipf.de!unlisys!news.snafu.de!home.ibert.com!news On Tue, 08 Dec 1998 18:09:47 GMT, alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) wrote: >And while we are at it, how in "hierarchical" Unix file system one can determine >the (used) size of "/" partition (excluding other file systems hanging off /) ? statfs? vstatfs? -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:52:34 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 18 Message-ID: <3681cb9a.15103630@news.home.ibert.com> References: <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> <36728afd.10991534@news.uunet.be> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On Tue, 15 Dec 1998 00:19:50 GMT, alderson@netcom.netcom.com (Richard M. Alderson III) wrote: >That's an awful lot of trouble to go to just to save a minuscule amount of disk >space, especially since the OS isn't going to let you umount /usr and /etc >while you're running. The only way what was originally suggested makes any >sense at all is if you *could* write to unmounted filesystems. Anything else >is just schratzing around for the sake of schratzing around. I can see another reason why you'd like to do that: to be able to mount / read-only. You cannot expect to run in production with all files in /etc read-only, but having a read-only root in production has one hell of a lot of reasons going for it, so having /etc on a separate filesystem makes a _lot_ of sense to me. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 20:55:47 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 16 Message-ID: <3682cc8c.15345669@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74nr8t$1si$2@news.latrobe.edu.au> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On 10 Dec 1998 06:55:25 GMT, cchd@lucifer.its.latrobe.edu.au (Huw Davies) wrote: >This property is vital if you run a mixed architecture VMS cluster where >you need to setup logicals pointing to the correct architecture executable >directory depending on the node being employed. Huh? Do I not remember being told in this very thread that an executable as old as myself will run on the latest greatest iron without as much as a hiccup? (Okay, two-thirds as old as myself.) -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 21:07:57 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 15 Message-ID: <3683cf52.16055441@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!195.252.142.12!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On Thu, 10 Dec 1998 21:14:36 -0600, hshubs@mindspring.com (Howard S Shubs) wrote: >The main difference here between UNIX and VMS, as I see it, is that files >on VMS are strongly typed. File attributes beyond size and dates are held >by the file header. While UNIX has some file attributes (I'm not real >solid on this), there appear to be relatively few of them, and they are >very general. Concerning the contents of the file, there are none. None at all. A file is a stream of bytes. Period. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 21:23:26 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 46 Message-ID: <3684d146.16555558@news.home.ibert.com> References: <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74nqnu$1si$1@news.latrobe.edu.au> <1djtoy1.a1f9ypmi4mj7N@n241-84.berlin.snafu.de> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.pop.de!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On Thu, 10 Dec 1998 18:52:47 +0100, jnickelsen@acm.org (Juergen Nickelsen) wrote: > - are directory entries of files; each file in a directory has at > least one Files do not live in directories, links do. Files live in filesystems. What about Links (1) - are directory entries of files; each file in a filesystem has at least one (2) Directories - are collections of links (1), and nothing more. Files in a filesystem - have an information block (called, for no especially good reason, an "i-node" or index node) that holds all their meta information like owner, size, access information, etc. - do not have a name as such, but ... - have at least one (2) link (1), which is an entry in a directory linking (hence the name) a file name to an actual file. - get one link (1) when they are created because you have to name them to create them > - the original directory entry of a file and additional ones are not > of a different kind (in other words, the first link (1) is not special in any way) Footnotes: (1) sometimes called "hard links" in reverse analogy to "soft links" (2) if this does not hold true, the filesystem is inconsistent >(corrections and additions welcome) I have added some technicalities to your excellent summary. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 21:57:45 -0600 Organization: MindSpring Enterprises Lines: 107 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> NNTP-Posting-Host: d1.56.d0.f5 X-Server-Date: 16 Dec 1998 03:57:49 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <3677bac7.10795833@news.home.ibert.com>, martin@ibert.com (Martin Ibert) wrote: >On Mon, 07 Dec 1998 10:07:15 -0600, hshubs@mindspring.com (Howard S >Shubs) wrote: >>2) lack of developer support other than the "make" command > >But that is something and blows MMS out of the water without even >trying. I can't argue. I've not yet managed to get MMS to do anything useful. >>3) lack of portability > >I don't think I can follow you there. I can port a program I have >written for any UN*Xoid (especially if I am sort of experienced and >write it in such a way to avoid very special features the UN*Xoid in >question might have) to any other UN*Xoid with little effort. But >where can I port a VMS program I have written? To OpenVMS? To UNIX. Pretty easily, but it does take a little knowledge of both OSs. But I really wish things went the other way. It's all DEC's fault for keeping the VMS OS so tightly held. >>(2) Let me ask you about the UNIX source-level debuggers. So, if you get >>a crash, you can use the "core" file. Great. >Indeed. >[more rambling, much of what is quite true, deleted] >What I really like is Purify. Okay, it costs money (rather a lot of >it), but it is really worth it. It will find almost every >dynamic-memory-related problem in almost no time at all for you. Not all errors are vm related. I'm (usually) using FORTRAN, and there's no dynamic memory coded into the programs I have to update. There -could- be, but engineers, in my experience, are taught FORTRAN-77, which has no standard dynamic memory features, and they don't seem terribly interested in picking up later versions, or other languages. I'm kinda pushing them at my current assignment. They aren't moving very quickly, at least in part because they've not started any new FORTRAN-based programs since I've been there. All programs I've written from scratch have been in either PERL or C. >Big wonder, technically speaking. I do understand that you route for >NT because it is Dave Cutler's child, but from painful experience I >can tell you that NT is an absolute bitch to develop for. Nah, I don't root for any kind of Windows. Dave Cutler sold out, as far as I can tell. I'm sure he had his reasons. >- Not a trace of consistency in the Win32 API. You have to look up the >return codes for every fucking call. What is the return value for >error? NULL? FALSE? INVALID_HANDLE_VALUE? Non-Zero? Christ, that's a >lot easier even in OS/2. With UNIX, it is essentially a no-brainer, >once you get the general hang of it. -1 is error if the value is an >int, NULL if it is a pointer. Exceptions are lamentable, but fairly >rare, aberrations [can't think of any right now]. It's a no-brainer in VMS too. Success is always an odd value, with the best success code being STATUS % 8 == 1. That's consistant across all VMS I've ever seen, and I've been working with it since VMS 3. I wonder why MS didn't do someting similar? Even Apple does that kind of thing, with noErr being the standard success value. >- The most feauritis-infested, bulky, unstable, unpleasant, klugy >development environment ever. You don't like UN*X source-code >debuggers? Try Microsoft's. You'll love gdb after that. (At least >Microsoft Visual Nightmare doesn't take all night to crash.) I'm looking forward to this. My next assignment is likely to send me to both Oracle and MS school. I'm looking forward to the Oracle school, not so much to the brainwashing of MS... but I like being able to pay the rent. >I have even developed for VMS once, and I kinda hated that as well, >because C was very poorly supported back then. (Is it now?) The >compiler was non-standard. (errno that is not an lvalue. Brrr.) If you >looked up a system call in the orange wall, there were code examples >galore. BLISS, Assembler, COBOL, you name it, they had it. But I don't >recall ever seeing a VAX C example. Back when the Orange Wall was current, there were several books of sample code published, in the smaller orange binders which could be bent to make them easel-like. They were titled something like "Using VMS Library Calls From FORTRAN" ...C" ...BLISS" ...COBOL", etc. One of the basic things you have to know as a VMS programmer is the calling standard and the concept of descriptors. >String handling is a mess. Plus, >there wasn't a way to read as many bytes from a file as stat would >tell you the file holds. You can open in one way and get more. You >open it another way and get less. Messy. From C, interacting with VMS runtime and system libraries was/is difficult. Yes. They don't want null-terminated character strings very often. They wanted strings pointed to by descriptor. And a descriptor is something that strikes me as very foreign to UNIX. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 21:59:53 -0600 Organization: MindSpring Enterprises Lines: 20 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <366E208A.E2CCAB80@tnglwood.demon.co.uk> <74nr8t$1si$2@news.latrobe.edu.au> <3682cc8c.15345669@news.home.ibert.com> NNTP-Posting-Host: d1.56.d0.f5 X-Server-Date: 16 Dec 1998 03:59:54 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!howland.erols.net!netnews.com!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <3682cc8c.15345669@news.home.ibert.com>, martin@ibert.com (Martin Ibert) wrote: >On 10 Dec 1998 06:55:25 GMT, cchd@lucifer.its.latrobe.edu.au (Huw >Davies) wrote: > >>This property is vital if you run a mixed architecture VMS cluster where >>you need to setup logicals pointing to the correct architecture executable >>directory depending on the node being employed. > >Huh? Do I not remember being told in this very thread that an >executable as old as myself will run on the latest greatest iron >without as much as a hiccup? I said 20-year old VAX to current VAX, and ditto for Alpha. In an expanded form, it would read, "a binary from a 20-year old VAX would work on a current VAX, and a binary from an original Alpha would work on a current Alpha." -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 22:00:49 -0600 Organization: MindSpring Enterprises Lines: 11 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <74jpov$4iu$1@agate.berkeley.edu> <74mabs$rrl$1@agate.berkeley.edu> <367551c8.59800829@news.uunet.be> <3683cf52.16055441@news.home.ibert.com> NNTP-Posting-Host: d1.56.d0.f5 X-Server-Date: 16 Dec 1998 04:00:50 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <3683cf52.16055441@news.home.ibert.com>, martin@ibert.com (Martin Ibert) wrote: >Concerning the contents of the file, there are none. None at all. A >file is a stream of bytes. Period. Right. However, sometimes the contents of a file are a pointer to another file (a symlink), or something else. There just don't seem to be many of these exceptions. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 15 Dec 1998 22:03:15 -0600 Organization: MindSpring Enterprises Lines: 11 Message-ID: References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <7523bh$cm$1@daemonweed.reanimators.org> <367ec532.13463432@news.home.ibert.com> NNTP-Posting-Host: d1.56.d0.f5 X-Server-Date: 16 Dec 1998 04:03:16 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <367ec532.13463432@news.home.ibert.com>, martin@ibert.com (Martin Ibert) wrote: >My VMS exposure has been rather limited (thank God), but I still think >"o for owner, g for group and o fo... shit!" once in a while. I explained the symbolic method to my troubled user today, and I think he got it! The accountant may have understood the binary/octal method, but I've not quizzed her yet. She's kinda busy with year-end stuff. -- Howard S Shubs The Denim Adept ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: lisard@zetnet.co.uk's message of 15 Dec 1998 15:13:58 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <755ubm$quj$1@irk.zetnet.co.uk> Date: Tue, 15 Dec 1998 22:50:29 GMT Lines: 17 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!netnews.com!ix.netcom.com!netcom!alderson In article <755ubm$quj$1@irk.zetnet.co.uk> lisard@zetnet.co.uk writes: >Whereas the other alternative just passes in the mask to the application and >leaves it to handle the globbing. Great, it lets them find associated files; >but ill-behaved program will refuse to glob at all, and the only way round >that is to have the shell do it. Your prejudices are showing. No one writing a file-handling program that expects multi-file specifications in a system in which wildcard expansion is handled by system calls would ignore those system calls. And a Unix program that only expects a single filename won't do well with globbing, either, now will it? -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 00:09:05 GMT Organization: . Lines: 38 Message-ID: <3676d777.52685247@news.uunet.be> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> NNTP-Posting-Host: pool02b-194-7-147-112.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach korpela@islay.ssl.berkeley.edu (Eric J. Korpela) on 14 Dec 1998 17:42:50 GMT to alt.folklore.computers: > In article <367541b4.0@192.168.0.20>, Samael wrote: > >>chmod u+x,go-rwx filename > > > >I prefer > > > >right-click menu > >properties option > >choose what flags you want to set on/off > > Try changing the permission of every file on your system that starts with > 'a' and ends with 'e'. Or all the ".exe" files in a directory for that > matter. See you tomorrow. On all drives, or just a single one? Searching subdirectories, or not? All files matching a certain mask, or all files of a certain type, or the two combined, or just plain all files in a directory or subtree? {Modified | created | last accessed} {before | after} a date you specify, or between two dates, or less than n {days | months} ago? All files {larger | smaller} than a size you specify? Searching all matching files for specific contents, or not? [Ooh, how I'd have liked to use the word "grep" here - but I must admit _that_ would be a lie, at least for the standard win* search] -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 00:09:07 GMT Organization: . Lines: 62 Message-ID: <367af967.61373440@news.uunet.be> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> NNTP-Posting-Host: pool02b-194-7-147-112.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news-fra1.dfn.de!news-fra.maz.net!newsfeed.tli.de!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach Robert Billing on Tue, 15 Dec 1998 06:57:01 +0000 to alt.folklore.computers: > Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, Don't you think that applies to any system you use daily? Or maybe it can be rephrased a bit: Simple things in Unix/Linux/VMS are difficult for novices, but once you get the basics, most more extended things are relatively easy: logarithmic learning curve, you could say. Simple things in Windows are easy even for novices, but even when you know the basics, some more extended things remain impossible to do: exponential learning curve. Are there any *n*x's that come with a usage counter on the 'man' command? I'd be interested in the results (no, not enough to have all my commands logged and write a shell script to count the number of occurrances of 'man' ;-) Both (logging and the shell script) fall in the 'impossible' category in windows (the scripting host is a step in the right direction, but it's a dragon - which doesn't mean I didn't manage to write a few very useful scripts). On my computer [but I admit, as I've done before, that I don't boot Linux very much] 'man' is one of the most used commands (after daily bread like 'ls' and a few other two letter words). Then, man is octal - that means the switches I can't remember are invariably more than 8 screens down ;-) This is why I created a directory 'notes' under my home dir, where I keep brief descriptions (1 screen max, preferably much less) of the things I might need again: I never needed this kind of help under Windows. I almost agree Windoze sucks in a lot of things, but in a lot of aspects I prefer it over Linux (BTW, did I ever tell you I as good as refused to use Windows until '95 came out? Windows was something for people who didn't know the first thing about computers, plus it was way too instable.) When I need a program I haven't used for over a year in Windows, I fire it up and usually just seeing it brings back the memories of how it works - if not, there's normally a Help button or menu. And that includes the most "dangerous" programs: starting them up just opens a window, to make them destroy something you have to give further commands. Under Linux it sometimes takes minutes of ls-ing _for_ the docs (it didn't come with a man page: now was it a plain doc file, or html, or info? In what directory?), followed by minutes of searching _through_ the docs, before even thinking of starting it. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: David Scheidt Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 01:41:21 GMT Organization: EnterAct L.L.C. Turbo-Elite News Server Lines: 64 Message-ID: <757341$m8$1@eve.enteract.com> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <367af967.61373440@news.uunet.be> NNTP-Posting-Host: 207.229.143.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.enteract.com!news.enteract.com!dscheidt Luc Van der Veken wrote: : Also sprach Robert Billing on : Tue, 15 Dec 1998 06:57:01 +0000 to alt.folklore.computers: : > Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, : Don't you think that applies to any system you use daily? No. It doesn't. The Windows box I have on my desk at work, which is there only so I can read my email is not easy to use. It simply doesn't do what I want it to. I am forced to use MicroSloth Exchange for email, and have not yet figured out how to get exchange to take a piece of mail that I get once a week, pipe it through a simple awk script to create a couple of data files, which get inserted into a Sybase database on a machine in another state, or not, depending on the results of some other processes that run on yet a third machine. I am forced to take the mail, save it manually, massage it a bit, and then run it through my awk scripts. This costs me 15 or 20 minutes a week, and annoys the hell out of me. : Are there any *n*x's that come with a usage counter on the 'man' : command? : I'd be interested in the results (no, not enough to have all my : commands logged and write a shell script to count the number of : occurrances of 'man' ;-) Just about all of them. Look at your system accounting sub-system. Tracking system resource usage is really handy thing, which is why hooks for it are built in most 'nixs. (I suspect it is one things required for branding but don't remember.) : Then, man is octal - that means the switches I can't remember are : invariably more than 8 screens down ;-) man(1) is by no means the most useful help system for a new user, assuming they manage to find it. It took me probably an hour to figure it out the very first time I used a Unix machine, sometime long ago. (Interestingly to the thread's topic, I was trying to log someone out of a terminal to use a VAX/VMS system.) Well written man pages provide all sorts of information, much of which you don't need most of the time. It is still there, for the those times when you need it. This is totally unlike Windows help, which doesn't let you figure out how to deal with rare occurances. How do you change what external program exchange runs when you click a url? Mine runs the Shockwave installer, which I find less than useful behavior. : This is why I created a directory 'notes' under my home dir, : where I keep brief descriptions (1 screen max, preferably much : less) of the things I might need again: I never needed this kind : of help under Windows. No, the options to format, reinstall from scratch, don't take up much more than a couple of lines. : Under Linux it sometimes takes minutes of ls-ing _for_ the docs : (it didn't come with a man page: now was it a plain doc file, or : html, or info? In what directory?), followed by minutes of Not providing a man(1) page is stupid. info(1) is one of the things I really don't like about gnu stuff. David Scheidt -- Actually, I wasn't trying to troll at all, but suddenly discovered this fishing line in my hands. I'm still not sure I udnerstand it, but I think some sort of meta-something is going on. If you figure it out, let me know. -- Dave Hatunen ###### From: David Scheidt Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 01:49:43 GMT Organization: EnterAct L.L.C. Turbo-Elite News Server Lines: 22 Message-ID: <7573jn$m8$2@eve.enteract.com> References: <755ubm$quj$1@irk.zetnet.co.uk> NNTP-Posting-Host: 207.229.143.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.enteract.com!news.enteract.com!dscheidt Richard M. Alderson III wrote: : In article <755ubm$quj$1@irk.zetnet.co.uk> lisard@zetnet.co.uk writes: : Your prejudices are showing. No one writing a file-handling program that : expects multi-file specifications in a system in which wildcard expansion : is handled by system calls would ignore those system calls. You don't work with the developers I do. They managed to do the same thing wrong 350 times in a row, when the correct behavior was three lines of code provided to them at least twice. : And a Unix program that only expects a single filename won't do well with : globbing, either, now will it? Depends on what it does with its arguments. If it is done smartly, or even brain-deadly, it ignores the second to nth args. David Scheidt -- David Scheidt Large fibreglass fruits are much the same the world over. -- Vicki Parslow Stafford ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <2230.653T2739T10516280@sky.bus.com> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 21 Message-ID: Date: Wed, 16 Dec 1998 03:04:02 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913777442 205.166.146.8 (Tue, 15 Dec 1998 21:04:02 CDT) NNTP-Posting-Date: Tue, 15 Dec 1998 21:04:02 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!chippy.visi.com!news-out.mcia.org!news-out.visi.com!ptah.visi.com!not-for-mail In article <2230.653T2739T10516280@sky.bus.com>, Charlie Gibbs wrote: >A fascinating alternate view. Thanks for pointing that out. >You've illustrated two totally different forms of abstraction. >And I'd say that's one of the most convincing arguments yet >that a system should provide both textual and graphic interfaces. One of the best arguments for this was AmigaDOS. If you used it for a while, you quickly found that the interaction of the two modes was a powerful strength. I find that, being very verbal, I miss the GUI less than I miss the CLI for 99% of what I do - but I really wish I had a GUI for the other few tasks, and that's why I have a Macintosh. ;) -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <755ubm$quj$1@irk.zetnet.co.uk> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 20 Message-ID: <2uFd2.2296$WZ6.6110370@ptah.visi.com> Date: Wed, 16 Dec 1998 03:07:42 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913777662 205.166.146.8 (Tue, 15 Dec 1998 21:07:42 CDT) NNTP-Posting-Date: Tue, 15 Dec 1998 21:07:42 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!chippy.visi.com!news-out.mcia.org!news-out.visi.com!ptah.visi.com!not-for-mail In article , Marco S Hyman wrote: >Snicker! As someone once said "never underestimate the power of human >stupidity!" Just because a system call or library function exists >does NOT mean it will be used. Or have you never seen this code (or >something very like it) in random applications: > for ( i = 0; i < count; i++ ) { > dst[ i ] = src[ i ]; > } Not only that, but once I benchmarked a test program, and memcpy was 280 times faster on the Sun workstation involved. Not sure why. :) -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: dpeschel@u.washington.edu (D. Peschel) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 05:30:46 GMT Organization: University of Washington, Seattle Lines: 12 Message-ID: <757gi6$1504$1@nntp1.u.washington.edu> References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <3678c0d1.12342645@news.home.ibert.com> NNTP-Posting-Host: saul6.u.washington.edu X-Trace: nntp1.u.washington.edu 913786246 37892 (None) 140.142.17.35 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: dpeschel Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.berkeley.edu!news.u.washington.edu!dpeschel In article <3678c0d1.12342645@news.home.ibert.com>, Martin Ibert wrote: >Solaris is not an operating system (at least it wasn't before Solaris >7). I think they call it an "operating environment" - it contains an >operating system (SunOS 5 in the case of Solaris 2) plus extra bits >and pieces. Is there such a thing as an anti-bit? If so, Solaris contains the compiler and development-library anti-bit most of the time, right? -- Derek ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <74jjnt$i01$1@vixen.cso.uiuc.edu> <74mmn1$4l7$3@irk.zetnet.co.uk> <367cc314.12921262@news.home.ibert.com> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 26 Message-ID: <6rId2.2322$WZ6.6197268@ptah.visi.com> Date: Wed, 16 Dec 1998 06:29:22 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913789762 205.166.146.8 (Wed, 16 Dec 1998 00:29:22 CDT) NNTP-Posting-Date: Wed, 16 Dec 1998 00:29:22 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <367cc314.12921262@news.home.ibert.com>, Martin Ibert wrote: >On 9 Dec 1998 20:31:29 GMT, lisard@zetnet.co.uk wrote: >>(The Bourne shell does have some advantages; it's a fairly lightweight >>scripting platform, and it keeps out of the way as much as possible. But >>I still don't think it's actually meant to be *used* by anyone. ;> ) >It is. For running scripts. :-) >The C shell, on the other hand, is not. I'd just like to point out that on NetBSD-current, and BSD/OS 3.x/4.x, '/bin/sh' is pretty useful as an interactive shell. Over time, it's acquired most of the "POSIX shell" features, as it should. So, while I still use ksh a lot, I occasionally migrate an account to plain old sh, and notice no significant differences. Just the ability to put $? in my prompt and have it evaluate has made the sh family a *huge* win for me. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: fleetvis@ricochet.net (phil) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 07:17:15 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 37 Message-ID: <36775b0a.17927949@news.ricochet.net> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> NNTP-Posting-Host: 204.179.131.77 X-Trace: 913792056 F7W8ZAFJM834DCCB3C usenet54.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!205.252.116.205!howland.erols.net!Supernews73!supernews.com!Supernews69!not-for-mail On Tue, 15 Dec 1998 19:57:14 GMT, martin@ibert.com (Martin Ibert) wrote: >On Mon, 07 Dec 1998 10:07:15 -0600, hshubs@mindspring.com (Howard S >Shubs) wrote: ,snip> >>UNIX... no wonder it's being supplanted by Windows NT. >Big wonder, technically speaking. I do understand that you route for >NT because it is Dave Cutler's child, but from painful experience I >can tell you that NT is an absolute bitch to develop for. I don't agree. A GUI is a bitch to develop for. Of all those available, NT is probably the easiest (you can produce a console app with minimal grief). >- Not a trace of consistency in the Win32 API. You have to look up the >return codes for every fucking call. What is the return value for >error? NULL? FALSE? INVALID_HANDLE_VALUE? Non-Zero? Christ, that's a >lot easier even in OS/2. With UNIX, it is essentially a no-brainer, >once you get the general hang of it. -1 is error if the value is an >int, NULL if it is a pointer. Exceptions are lamentable, but fairly >rare, aberrations [can't think of any right now]. Yes. This really sucks. There's boolean, integer, pointer returns etc. and they all seem to vary in failure rules. >- The most feauritis-infested, bulky, unstable, unpleasant, klugy >development environment ever. You don't like UN*X source-code >debuggers? Try Microsoft's. You'll love gdb after that. (At least >Microsoft Visual Nightmare doesn't take all night to crash.) Now you have to be joking. The MSVC debugger is way beyond anything that i've ever previously used. I really doubt that it was developed by Microsoft, it's that good. Our embedded guys (including Linux hackers) drool when they see it. Combine it with Bounds Checker and you have the most proficient debugging environment short of using an interpreted language. phil. ###### From: fleetvis@ricochet.net (phil) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 07:34:22 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 22 Message-ID: <36775f84.19073936@news.ricochet.net> References: <755ubm$quj$1@irk.zetnet.co.uk> NNTP-Posting-Host: 204.179.131.77 X-Trace: 913793490 F7W8ZAFJM834DCCB3C usenet78.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail On 15 Dec 1998 15:13:58 GMT, lisard@zetnet.co.uk wrote: >Hell, they both have good points and bad points, don't they? In Unix, no >program ever has to worry about expanding filenames, because it's done >for them; all the program has to do is go down its list of files and do >things with each. The globbing might not always do what you want, but >most of the time it works pretty well. > >Whereas the other alternative just passes in the mask to the application >and leaves it to handle the globbing. Great, it lets them find >associated files; but ill-behaved program will refuse to glob at all, >and the only way round that is to have the shell do it. OK, if i write a Unix app, can i gain access to the original command line? If that is possible, then i suppose the app can make the decision on whether to use the expanded command line or not and rm * becomes a controllable activity. It should be perfectly possible to distinguish between such a command in a script as opposed to it being entered on a command line (and to take the appropriate action or supply the appropriate user interface). phil. ###### From: Ariel Scolnicov Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 10:35:19 +0200 Organization: NetVision Israel Lines: 48 Message-ID: References: <755ubm$quj$1@irk.zetnet.co.uk> <36775f84.19073936@news.ricochet.net> NNTP-Posting-Host: bioserv.compugen.co.il X-Trace: news.netvision.net.il 913797401 2846 194.90.227.153 (16 Dec 1998 08:36:41 GMT) X-Complaints-To: abuse@netvision.net.il NNTP-Posting-Date: 16 Dec 1998 08:36:41 GMT X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer.gip.net!news-dc.gip.net!news.gsl.net!gip.net!news-feed.netvision.net.il!194.90.1.15.MISMATCH!news!not-for-mail fleetvis@ricochet.net (phil) writes: > On 15 Dec 1998 15:13:58 GMT, lisard@zetnet.co.uk wrote: > >Hell, they both have good points and bad points, don't they? In Unix, no > >program ever has to worry about expanding filenames, because it's done > >for them; all the program has to do is go down its list of files and do > >things with each. The globbing might not always do what you want, but > >most of the time it works pretty well. > > > >Whereas the other alternative just passes in the mask to the application > >and leaves it to handle the globbing. Great, it lets them find > >associated files; but ill-behaved program will refuse to glob at all, > >and the only way round that is to have the shell do it. > > OK, if i write a Unix app, can i gain access to the original command > line? If that is possible, then i suppose the app can make the > decision on whether to use the expanded command line or not and rm * > becomes a controllable activity. It should be perfectly possible to > distinguish between such a command in a script as opposed to it being > entered on a command line (and to take the appropriate action or > supply the appropriate user interface). No. There's no such thing as "the original command line". Programs get executed when somebody makes an exec..() call, passing arguments. Definitely no globbing there. Shells, as a convenience to the user, glob filenames, and only then make the exec..() call (with the appropriate args). So "the original command line" doesn't exist. On the other hand, only last week someone noticed a program behaving in an "impossible" manner. She had several files ending in _S, but none ending in _s. Recall (from this thread) that Un*x is case-sensitive. But "program ... *_s" was catching the _S files! (Originally she'd had files of each type, and wanted to run only on one type). It turned out that in the interests of cross-platform compatibility (i.e. it doesn't work, equally well, on all platforms) the program was performing its own globbing (and working in a somewhat case-insensitive manner), and "program" was appropriately aliased (in the "standard environment") to "set noglob" and then run "program" (the program) with these args. Which is why we call them "(Aaaa)arg(h)s". -- Ariel Scolnicov /---------------\ "GCAAGAATTGAACTGTAG" Compugen Ltd. |Join the Dark | Tel: +972-2-6795059 (Jerusalem) 17 Hamacabim St. |Side! Just Say:| Tel: +972-3-9348482 (Main office) Petakh-Tikva, 49220 | "Yes, but..."| Fax: +972-3-9348489 ISRAEL \---------------/ ariels@compugen.co.il ###### From: Ariel Scolnicov Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 10:51:43 +0200 Organization: NetVision Israel Lines: 68 Message-ID: References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> NNTP-Posting-Host: bioserv.compugen.co.il X-Trace: news.netvision.net.il 913798384 2846 194.90.227.153 (16 Dec 1998 08:53:04 GMT) X-Complaints-To: abuse@netvision.net.il NNTP-Posting-Date: 16 Dec 1998 08:53:04 GMT X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news-peer.gip.net!news-dc.gip.net!news.gsl.net!gip.net!news-feed.netvision.net.il!194.90.1.15.MISMATCH!news!not-for-mail Mark Wilkins writes: > [...] > > I work with a number of people who have essentially an artistic > background and have learned to use computers. I also work with a number > of people from a software engineering background. > > A key difference between these two groups of people is that the > artists have a much stronger memory for spatial relationships and a much > weaker memory for verbal syntaxes. Remembering which of two commands to > use in a command-line environment is a verbal game, essentially, while > remembering which command serves a given function in a graphic user > interface is a matter of spatial memory. > > My point is that the artists often find graphic menus far easier to > use even after they've achieved a high level of proficiency with a > software package because the means of interaction is suited to their > strengths. Meanwhile, software engineers find themselves more > comfortable with command lines and often ask why graphic interfaces are > even necessary. Many of the artists require repeated explanations to > remember a textual command but only need to be shown a menu item once. > > What is a fallacy is to suggest that to have difficulty with textual > commands is equivalent to failure to learn the software or achieve the > state of "knowing what you're doing." Though these artists may not have > the same strengths in symbolic manipulation that I do, they are able to > use their spatial memory to be productive in ways I could never dream > of. > Nobody is disputing the ease of using mice to perform simple tasks. This will probably offend GUI people, but GUIs have an extremely limited syntax, essentially "verb object" (sometimes "subject verb object"), but only for simple verbs and objects (typically the only objects are nouns or one of the magic nouns "clipboard" or "selection". If you like, the GUI user is saying "OG WANT COMPUTER DO THIS". That's not enough for performing slightly more complex files, like "delete all large .avi files". Now, as has been pointed out on this thread (by "Samael", at least), you can break this up into "select all large .avi files" (which will probably involve at least some CLI stuff, e.g. you'll need to use the abstraction "*.avi", and some hybrid graphical/verbal "AND" operator) and then "delete selection". But you've already doing some of what the computer should be doing -- translating a high-level concept into the corresponding low-level sequence of operations. To be fair, the correct graphical analogue would be "delete all files with icons on the bottom-right of the screen [which is where I (hypothetically) store my large .avi files]", but there are only so many points on the compass, and this would require much advance planning and discipline, so my objection stands. Now let's say I managed to say "delete all large .avi files". Now I want to say "delete all large .avi files every day at midnight". Notice how I can recursively create a new sentence by adding "every day at midnight" to a correct command. Visual manipulation just can't handle the abstraction. OG NO SEE BIG FILES WHEN DARK. >[...] > mark_wilkins @fa.disney.com Burbank, CA -- Ariel Scolnicov /---------------\ "GCAAGAATTGAACTGTAG" Compugen Ltd. |Join the Dark | Tel: +972-2-6795059 (Jerusalem) 17 Hamacabim St. |Side! Just Say:| Tel: +972-3-9348482 (Main office) Petakh-Tikva, 49220 | "Yes, but..."| Fax: +972-3-9348489 ISRAEL \---------------/ ariels@compugen.co.il ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 11:10:04 -0000 Lines: 31 Message-ID: <36779525.0@192.168.0.20> References: <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <36767e0a.0@192.168.0.20> <1998Dec15.190813.10619@picarefy.picarefy.com> NNTP-Posting-Host: 195.147.140.145 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 James W. Birdsall wrote in message <1998Dec15.190813.10619@picarefy.picarefy.com>... >In article <36767e0a.0@192.168.0.20> "Samael" writes: >>It's not a speed thing, it's a memory thing. If I want to move some text >>from one place t another, it's erasy to look at the menus, spot '"Edit' and >>then look down the listun til I see the word 'Cut' and think, aah, that's >>what I want, than try and remember any key combinations. > > If you have to remember key combinations, you haven't learned the >program yet. When you have, as soon as you know what you want to do, your >hands will do it, in the same way that you type text without thinking about >hitting the individual keys. > This is why so many people want keyboard shortcuts. Menus *are* handy for >infrequently-used functions, but being forced to go through them just slows >down experienced users. Certianly do. Which is why I use the keyboard for most things. But most people aren't experts. Computer are being used by non-experts more and more. I've not found many things I couldn't do in windows using the keyboard (fpor batch jobs I tend to use the Command Window/DOS). Samael ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 98 11:24:42 GMT Organization: UltraNet Communications, Inc. Lines: 21 Message-ID: <7586n2$epi$1@ligarius.ultra.net> References: <755ubm$quj$1@irk.zetnet.co.uk> <7573jn$m8$2@eve.enteract.com> NNTP-Posting-Host: d1.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 16 Dec 1998 11:48:50 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!feed1.news.rcn.net!rcn!news.ultranet.com!d1 In article <7573jn$m8$2@eve.enteract.com>, David Scheidt wrote: >Richard M. Alderson III wrote: >: In article <755ubm$quj$1@irk.zetnet.co.uk> lisard@zetnet.co.uk writes: >: Your prejudices are showing. No one writing a file-handling >: program that expects multi-file specifications in a system >: in which wildcard expansion >: is handled by system calls would ignore those system calls. > >You don't work with the developers I do. They managed to >do the same thing wrong 350 times in a row, when the correct >behavior was three lines of code provided to them at least twice. ROTFLMAO. Comfort, comfort, comfort. In my group, we wouldn't call them developers :-))))) This one's going to keep me chuckling all day. Thanks. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 98 11:40:37 GMT Organization: UltraNet Communications, Inc. Lines: 29 Message-ID: <7587kt$epi$2@ligarius.ultra.net> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <367af967.61373440@news.uunet.be> NNTP-Posting-Host: d1.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 16 Dec 1998 12:04:45 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.xcom.net!news.ultranet.com!d1 In article <367af967.61373440@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach Robert Billing on >Tue, 15 Dec 1998 06:57:01 +0000 to alt.folklore.computers: > >> Exactly. Unix/Linux/VMS *looks* difficult to use, but is actually easy, > >Don't you think that applies to any system you use daily? > >Or maybe it can be rephrased a bit: > >Simple things in Unix/Linux/VMS are difficult for novices, but >once you get the basics, most more extended things are relatively >easy: logarithmic learning curve, you could say. > >Simple things in Windows are easy even for novices, but even when >you know the basics, some more extended things remain impossible >to do: exponential learning curve. Errmmm....not for this novice. I found that attempting to find out how to do the simple things in Windows was impossible. It kept deciding what it thought I wanted to do. My learning curve was completely involved in learning how to undo what I didn't want it to do. /BAH Subtract a hundred and four for e-mail. ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 12:16:40 -0000 Lines: 89 Message-ID: <3677a4c1.0@192.168.0.20> References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> NNTP-Posting-Host: 195.147.140.18 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 Ariel Scolnicov wrote in message ... >Mark Wilkins writes: > >> [...] >> >> I work with a number of people who have essentially an artistic >> background and have learned to use computers. I also work with a number >> of people from a software engineering background. >> >> A key difference between these two groups of people is that the >> artists have a much stronger memory for spatial relationships and a much >> weaker memory for verbal syntaxes. Remembering which of two commands to >> use in a command-line environment is a verbal game, essentially, while >> remembering which command serves a given function in a graphic user >> interface is a matter of spatial memory. >> >> My point is that the artists often find graphic menus far easier to >> use even after they've achieved a high level of proficiency with a >> software package because the means of interaction is suited to their >> strengths. Meanwhile, software engineers find themselves more >> comfortable with command lines and often ask why graphic interfaces are >> even necessary. Many of the artists require repeated explanations to >> remember a textual command but only need to be shown a menu item once. >> >> What is a fallacy is to suggest that to have difficulty with textual >> commands is equivalent to failure to learn the software or achieve the >> state of "knowing what you're doing." Though these artists may not have >> the same strengths in symbolic manipulation that I do, they are able to >> use their spatial memory to be productive in ways I could never dream >> of. >> > >Nobody is disputing the ease of using mice to perform simple tasks. >This will probably offend GUI people, but GUIs have an extremely >limited syntax, essentially "verb object" (sometimes "subject verb >object"), but only for simple verbs and objects (typically the only >objects are nouns or one of the magic nouns "clipboard" or >"selection". If you like, the GUI user is saying "OG WANT COMPUTER DO >THIS". > >That's not enough for performing slightly more complex files, like >"delete all large .avi files". Now, as has been pointed out on this >thread (by "Samael", at least), you can break this up into "select all >large .avi files" (which will probably involve at least some CLI >stuff, e.g. you'll need to use the abstraction "*.avi", and some >hybrid graphical/verbal "AND" operator) and then "delete selection". Just two different boxes to enter info into. Admittedly, it makes life tricky if the programmer hasn't anticipated you wanting that functionality. >But you've already doing some of what the computer should be doing -- >translating a high-level concept into the corresponding low-level >sequence of operations. To be fair, the correct graphical analogue >would be "delete all files with icons on the bottom-right of the >screen [which is where I (hypothetically) store my large .avi files]", >but there are only so many points on the compass, and this would >require much advance planning and discipline, so my objection stands. From the explorer, I can switch to 'detailed' view, which lists the files in columns with some of their attributes (including size and dates), click on the 'size' header, which sorts the files in size order. Go to the point I want to select from and then drag downwards from there. Not as efficient as command line, certainly, but a tad more intuitive. >Now let's say I managed to say "delete all large .avi files". Now I >want to say "delete all large .avi files every day at midnight". >Notice how I can recursively create a new sentence by adding "every >day at midnight" to a correct command. Visual manipulation just can't >handle the abstraction. OG NO SEE BIG FILES WHEN DARK. You are right, for that sort of thing we have to go back to a batch file and then call it from the Task Scheduler. A Gui can never be as felxible as a CLI, becasue (in theory) a CLI allows you to do _anything_, but for the subset of things it deals with, a GUI should be more intuitive and easier to use (unless you're the sort of person who doesn't think graphically). Samael ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 13:04:52 GMT Organization: Daedalus Consulting Lines: 44 Message-ID: <758b5k$db$1@news.wlg.netlink.net.nz> References: <73ela0$51h@hermes.acs.unt.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Martin Ibert wrote: >>2) lack of developer support other than the "make" command > >But that is something and blows MMS out of the water without even >trying. Huh? MMS *is* make, with a couple of VMSisms bolted on the side. >The changes will be absolutely marginal compared with the changes >needed to port your VMS program to _anywhere_. So I really don't see >your point at all. Porting from VMS to VMS or VMS to OpenVMS doesn't >count. Unix code, by which I mean stuff that uses more than the basic library stuff described in K&R for the most part only ports easily to unix systems. Quite a lot of unix stuff is emulated by C libraries on other, but tends to to come unstuck fairly quickly. Genuinely portable C code can be written on any OS. It just better not do anything tricky like expect any more than line by line terminal input or making TCP connections... >I have even developed for VMS once, and I kinda hated that as well, >because C was very poorly supported back then. (Is it now?) The C never was a good match for VMS. Mainly because VMS makes extensive use of descriptors for parameter passing, and all the libraries and system calls use descriptors or at least address and length pairs. Most languages under VMS pass string parameters using descriptors. C unfortunately, lacks any such concept -- oh sure, you can construct the things yourself, but there's no native support in C for a string with a length word, and the C library glosses over this inadequacy (for inadequacy it is, and I curse it every time yet another bloody buffer overflow bug is discovered) by the use of null-terminated strings. That, and a general lack of support from various camps within DEC -- things have changed now, but at one point the VAXCRTL development effort was less than half a person, and VAXCRTL was a piece of crap that seriously needed some real work done to it. It was down to such bogosities as memcpy going into an infinite loop if faced with more than 64k to copy... -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: slanning@buphy.bu.edu (Scott Lanning) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 15 Dec 1998 22:10:15 GMT Organization: Boston Univ. Physics Lines: 21 Message-ID: <756mo7$3o3$1@news1.bu.edu> References: <74tckl$al@axalotl.demon.co.uk> <74v2ap$pbc$1@news1.bu.edu> <367a92b5.1515509@news.uunet.be> NNTP-Posting-Host: buphy.bu.edu X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.eecs.umich.edu!news.bu.edu!slanning Luc Van der Veken (lucvdv@null.net) wrote: : Also sprach slanning@buphy.bu.edu (Scott Lanning) on 13 Dec 1998 : > For permission to read OR write, bitwise OR : > r = 100 : > w = 010 : > --------------- (mnemonic: ones are heavy, : > r | w = 110 so they drop through) : Not bad if you're explaining it to a programmer, but I think for : 'ordinary' people the confusion starts at the first line. : : In common speach 'OR' usually (or at least often) means exclusive You're right. In fact, when I wrote the above, I originally put "read AND write", and didn't include "bitwise OR". Only after I realized that (1 & 0 = 1) was wrong, did I change it. (For thought efficiency, I often "alias" my thoughts, then fill in details later, if necessary--it's a.k.a., "handwaving". :) If explaining to a non- programmer, I wouldn't mention bitwise OR; however, one of the previous posters said he'd been a programmer but just hadn't bothered to remember the bitwise interpretation. ###### Date: 16 Dec 98 11:12:42 -0800 From: "Charlie Gibbs" Subject: Re: Is VMS taboo, or what? References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <2230.653T2739T10516280@sky.bus.com> Message-ID: <1460.654T2115T6725648@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 33 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 16 Dec 1998 12:11:30 -0800, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news-in-east1.sprintlink.net!news.sprintlink.net!news.westel.com!news.skybus.com!204.244.247.104 In article seebs@plethora.net (Peter Seebach) writes: >In article <2230.653T2739T10516280@sky.bus.com>, >Charlie Gibbs wrote: > >>A fascinating alternate view. Thanks for pointing that out. >>You've illustrated two totally different forms of abstraction. >>And I'd say that's one of the most convincing arguments yet >>that a system should provide both textual and graphic interfaces. > >One of the best arguments for this was AmigaDOS. If you used it for >a while, you quickly found that the interaction of the two modes was a >powerful strength. What do you mean "was"? I'm using it this minute. And indeed, the smooth integration of GUI and CLI is one of its joys. >I find that, being very verbal, I miss the GUI less than I miss the CLI >for 99% of what I do - but I really wish I had a GUI for the other few >tasks, and that's why I have a Macintosh. ;) I too am more text-oriented. But like you I occasionally prefer to use a GUI. My Amigas might not have the latest flashy software, but they do what I need, with the flexibility that I want. And thanks to its lean, clean design, this Amiga with its 25-MHz 68030 feels in many ways much faster than a 200-MHz Pentium box running Windows. (For instance, the Amiga doesn't gronk the hard drive for 5 seconds just to unblank the screen...) -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: viro@artin.math.psu.edu (Alexander Viro) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 12:42:16 -0500 Organization: -ENOENT Lines: 28 Message-ID: <758rdo$12@artin.math.psu.edu> References: <749lip$gjf$1@vixen.cso.uiuc.edu> <74nqnu$1si$1@news.latrobe.edu.au> <1djtoy1.a1f9ypmi4mj7N@n241-84.berlin.snafu.de> <3684d146.16555558@news.home.ibert.com> NNTP-Posting-Host: artin.math.psu.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsserver.jvnc.net!news3.cac.psu.edu!not-for-mail In article <3684d146.16555558@news.home.ibert.com>, Martin Ibert wrote: >On Thu, 10 Dec 1998 18:52:47 +0100, jnickelsen@acm.org (Juergen >Nickelsen) wrote: > >> - are directory entries of files; each file in a directory has at >> least one > >Files do not live in directories, links do. Files live in filesystems. >What about > >Links (1) > >- are directory entries of files; each file in a filesystem has at >least one (2) [snip] >(2) if this does not hold true, the filesystem is inconsistent Bzzzert. open("foo",O_RDONLY); unlink("foo"); and there you go - file exists but is not accessible from the namespace. It will be freed when the last process will close it. Link count is bloody 0. Ditto for binary of running process (at least on systems with demand-loading). Ditto for a shared library. Link prevents file from being freed, but there may be other holders. -- "You're one of those condescending Unix computer users!" "Here's a nickel, kid. Get yourself a better computer" - Dilbert. ###### From: viro@riemann.math.psu.edu (Alexander Viro) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 14:38:48 -0500 Organization: -ENOENT Lines: 32 Message-ID: <759288$6vt@riemann.math.psu.edu> References: <73ela0$51h@hermes.acs.unt.edu> <366f6b01.9489096@Rockyd> <3680c9be.14628000@news.home.ibert.com> <36780619.96636558@Rockyd> NNTP-Posting-Host: riemann.math.psu.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsserver.jvnc.net!news3.cac.psu.edu!not-for-mail In article <36780619.96636558@Rockyd>, Alexandre Pechtchanski wrote: >On Tue, 15 Dec 1998 20:43:02 GMT, martin@ibert.com (Martin Ibert) wrote: >[ Courtesy cc'ed through e-mail to the quoted author ] > >>On Tue, 08 Dec 1998 18:09:47 GMT, alex*@*rockvax.rockefeller.edu >>(Alexandre Pechtchanski) wrote: >> >>>And while we are at it, how in "hierarchical" Unix file system one can determine >>>the (used) size of "/" partition (excluding other file systems hanging off /) ? >> >>statfs? vstatfs? > >clinfo> uname -a >ULTRIX clinfo.rockefeller.edu 4.3 1 RISC >clinfo> statfs >statfs: not found >clinfo> find / -name statfs -print >clinfo> > >Same with vstatfs. Any other ideas? > >[ When replying, remove *'s from address ] >Alexandre Pechtchanski, Systems Manager, RUH, NY Those are syscalls. Details of required man invocation are left as an exercise for reader. -- "You're one of those condescending Unix computer users!" "Here's a nickel, kid. Get yourself a better computer" - Dilbert. ###### Sender: marc@dumbcat.snafu.org Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> From: Marco S Hyman Date: 16 Dec 1998 15:16:35 -0800 Message-ID: Organization: S.N.A.F.U. (www.snafu.org) X-Newsreader: Gnus v5.5/Emacs 20.3 Lines: 24 NNTP-Posting-Host: 204.94.187.130 X-Trace: nntp1.ba.best.com 913850198 214 marc@204.94.187.130 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.berkeley.edu!news2.best.com!news3.best.com!nntp1.ba.best.com!not-for-mail alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) writes: > And while we are at it, how in "hierarchical" Unix file system one can > determine the (used) size of "/" partition (excluding other file systems > hanging off /) ? See df(1). Output on my system looks something like: Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/sd0a 36655 23930 10893 69% / /dev/sd0d 36655 21430 13393 62% /var /dev/sd0e 293359 208988 69704 75% /usr /dev/sd3g 242887 162263 68480 70% /usr/local /dev/sd0f 586735 298651 258748 54% /usr1 /dev/sd3h 1537981 1104171 280012 80% /usr2 /dev/sd2g 906513 193894 621968 24% /cvs /dev/sd0g 900615 588045 267540 69% /openbsd kernfs 97984 8308 0 100% /kern mfs:11560 126553 37 120189 0% /tmp The column labeled "Used" contains the info you're looking for. // marc ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 17:35:09 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 27 Message-ID: <367760df.55362895@news.vip.net> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <7523bh$cm$1@daemonweed.reanimators.org> <367ec532.13463432@news.home.ibert.com> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.80 X-Trace: 913829152 A01OARAUVD450CCD1C usenet54.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!Supernews73!supernews.com!Supernews69!not-for-mail hshubs@mindspring.com (Howard S Shubs) wrote: >In article <367ec532.13463432@news.home.ibert.com>, martin@ibert.com >(Martin Ibert) wrote: > >>My VMS exposure has been rather limited (thank God), but I still think >>"o for owner, g for group and o fo... shit!" once in a while. > >I explained the symbolic method to my troubled user today, and I think he >got it! The accountant may have understood the binary/octal method, but >I've not quizzed her yet. She's kinda busy with year-end stuff. If she does it by hand, what is the probability that she does it in octal? Apparently, the late Rear Admiral Grace Hopper during one period of time messed up her che[ck|que]book by doing it in a mix of octal and decimal. Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### From: alex*@*rockvax.rockefeller.edu (Alexandre Pechtchanski) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Organization: Rockefeller University Hospital (GCRC), New York Message-ID: <36780619.96636558@Rockyd> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> <3680c9be.14628000@news.home.ibert.com> X-Newsreader: Forte Agent 1.5/32.451 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 22 Date: Wed, 16 Dec 1998 19:17:16 GMT NNTP-Posting-Host: 129.85.24.56 X-Trace: rockyd.rockefeller.edu 913835843 129.85.24.56 (Wed, 16 Dec 1998 14:17:23 EDT) NNTP-Posting-Date: Wed, 16 Dec 1998 14:17:23 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.nyu.edu!rockyd.rockefeller.edu!not-for-mail On Tue, 15 Dec 1998 20:43:02 GMT, martin@ibert.com (Martin Ibert) wrote: [ Courtesy cc'ed through e-mail to the quoted author ] >On Tue, 08 Dec 1998 18:09:47 GMT, alex*@*rockvax.rockefeller.edu >(Alexandre Pechtchanski) wrote: > >>And while we are at it, how in "hierarchical" Unix file system one can determine >>the (used) size of "/" partition (excluding other file systems hanging off /) ? > >statfs? vstatfs? clinfo> uname -a ULTRIX clinfo.rockefeller.edu 4.3 1 RISC clinfo> statfs statfs: not found clinfo> find / -name statfs -print clinfo> Same with vstatfs. Any other ideas? [ When replying, remove *'s from address ] Alexandre Pechtchanski, Systems Manager, RUH, NY ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 20:42:02 GMT Organization: . Lines: 22 Message-ID: <36881af5.12697427@news.uunet.be> References: <1456.648T1620T6105137@sky.bus.com> <74qpdt$57u@axalotl.demon.co.uk> <7523bh$cm$1@daemonweed.reanimators.org> <367ec532.13463432@news.home.ibert.com> NNTP-Posting-Host: pool02b-194-7-148-23.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach martin@ibert.com (Martin Ibert) on Tue, 15 Dec 1998 20:24:21 GMT to alt.folklore.computers: > On 13 Dec 1998 20:14:41 -0800, Frank McConnell > wrote: > > >It confuses folks like me who have had enough exposure to VMS that we > >have to sit there and think about whether "o" means "owner" like it > >does in SET FILE/PROT. Generally speaking, by the time I remember I > >worked out what I want in octal anyway. > > My VMS exposure has been rather limited (thank God), but I still think > "o for owner, g for group and o fo... shit!" once in a while. I've never been exposed to VMS, yet I've got similar rememberings - and I guess I'm not the only one who thought, upon reading Frank's post, "wouldn't it be a good time to go checking the permissions on some of my files?" -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 20:42:05 GMT Organization: . Lines: 36 Message-ID: <3679edd7.1148000@news.uunet.be> References: <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <36767e0a.0@192.168.0.20> <1998Dec15.190813.10619@picarefy.picarefy.com> <36779525.0@192.168.0.20> NNTP-Posting-Host: pool02b-194-7-148-23.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach "Samael" on Wed, 16 Dec 1998 11:10:04 -0000 to alt.folklore.computers: > > James W. Birdsall wrote in message > >This is why so many people want keyboard shortcuts. Menus *are* handy for > >infrequently-used functions, but being forced to go through them just slows > >down experienced users. > > Certianly do. Which is why I use the keyboard for most things. Me too But depending somewhat on what program it is: some are less keyboard friendly than they ought to be. [AAMOF, there may be a key combination for it, but this afternoon I grabbed my mouse to center a paragraph in Word because ctrl/c didn't do the job] > But most > people aren't experts. Computer are being used by non-experts more and > more. True, but OTOH that sounds like what the experts said 20 years ago, when I was a novice. > I've not found many things I couldn't do in windows using the > keyboard (fpor batch jobs I tend to use the Command Window/DOS). Usually the same here, but sometimes I don't even go that far: doubleclick to run one, rightclick/edit to edit one, rightclick/new/textdoc to create a new one. What more does one need (assuming it's a bat that doesn't need arguments)? -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 20:42:08 GMT Organization: . Lines: 83 Message-ID: <36831906.12202596@news.uunet.be> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <367af967.61373440@news.uunet.be> <757341$m8$1@eve.enteract.com> NNTP-Posting-Host: pool02b-194-7-148-23.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach David Scheidt on 16 Dec 1998 01:41:21 GMT to alt.folklore.computers: > No. It doesn't. The Windows box I have on my desk at work, which is there > only so I can read my email is not easy to use. It simply doesn't do what > I want it to. I suppose you mean it doesn't interprete your commands the way you are accustomed to. Comfort: true the other way around too. > I am forced to use MicroSloth Exchange for email, and have > not yet figured out how to get exchange to take a piece of mail that I get > once a week, pipe it through a simple awk script to create a couple of data > files, which get inserted into a Sybase database on a machine in another state, or not, depending on the results of some other processes that run on yet a > third machine. I am forced to take the mail, save it manually, massage it a > bit, and then run it through my awk scripts. This costs me 15 or 20 minutes > a week, and annoys the hell out of me. Are you expecting that you can start doing things that required you to learn shell scripts and awk and do a lot of tinkering on unix, on another system right from the start - just because you know how to do them on unix? Experienced Windows users face a slope at least as steep when they step into unix (I think even a lot steeper). Awk for example is typical for unix (at least I think it is) - if you want to do the same thing under windows, the way MS would probably like you most to do it is to open the message in Word and write a Word macro to do the work. The awk script would be shorter, but then again if you wanted to do something entirely different you'd have to switch to another scripting language. Under windows you stay with the same: the [dis]advantages of general purpose languages. > need it. This is totally unlike Windows help, which doesn't let you figure > out how to deal with rare occurances. How do you change what external program > exchange runs when you click a url? Mine runs the Shockwave installer, which > I find less than useful behavior. Treat it as if it were a file (this _is_ documented, but AFAIR they don't tell you it works for other things as files. Works for disks and directories too, btw). Like for any file type, you change how it's opened from the shell (sorry, I meant explorer). View / folder options, tab "file types". [I agree that's not the most obvious place for that menu, OK?] Find the right type of URL (if you meant a HTML document: it's called just that in the list, if not: you can find all kinds of URL's under the letter U), click edit. Probably all it takes is setting another action as the default. If not, select the type of action you want to change (you'll want the one in bold print, = default, now) and click edit again, or create an all new action. Enter the command line you want the system to use. Or, you can solve it like most users do: reformat your harddisk, and reinstall Windows. If you can get [a shortcut to] the same kind of url in an explorer window somehow, it might be easier [assuming this works on url's the same as on other files - I never tried]: select it, then shift/right click it [again, I agree it's everything but obvious why you should first select it, but if you don't you get the wrong menu], select "open with", point to the right program (IE4 or something), check "always use this program...", done. > No, the options to format, reinstall from scratch, don't take up much more than > a couple of lines. None at all, actually. I fdisked & reinstalled my Linux partitions often enough before I had an acceptable layout to know everything about that without help :-) [I've got 2 HD's, with Linux installed in multiple partitions divided over the 2: an attempt to minimize head movement, in several movements.] -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 20:42:10 GMT Organization: . Lines: 43 Message-ID: <367d0f35.9689933@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> NNTP-Posting-Host: pool02b-194-7-148-23.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news-peer1.sprintlink.net!news.sprintlink.net!news.maxwell.syr.edu!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach fleetvis@ricochet.net (phil) on Wed, 16 Dec 1998 07:17:15 GMT to alt.folklore.computers: > On Tue, 15 Dec 1998 19:57:14 GMT, martin@ibert.com (Martin Ibert) > wrote: > >- Not a trace of consistency in the Win32 API. You have to look up the > >return codes for every fucking call. What is the return value for > >error? NULL? FALSE? INVALID_HANDLE_VALUE? Non-Zero? Christ, that's a > >lot easier even in OS/2. With UNIX, it is essentially a no-brainer, > >once you get the general hang of it. -1 is error if the value is an > >int, NULL if it is a pointer. Exceptions are lamentable, but fairly > >rare, aberrations [can't think of any right now]. > Yes. This really sucks. There's boolean, integer, pointer returns etc. > and they all seem to vary in failure rules. I always thought they were pretty straight-forward. If the function is supposed to return a "something", 0 means an error (in a broad sense: NULL==0). If it's returning a status code, 0 means success and a non-zero return value can be examined more closely to find the reason of the error (quite like exit codes of programs - an approach I think it borrowed from that other OS where C also originates from). If it's boolean, I guess I don't have to say what indicates failure. And all those fancy names like INVALID_HANDLE_VALUE and STATUS_SUCCESS are just macros defined as [see above] in the right data type, intended to make your programs more readable && portable. >> > <[MSVC debugger + BoundsChecker] best short of interpreted env> Agree, and I have yet to experience its first crash. -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Wed, 16 Dec 1998 20:42:13 GMT Organization: . Lines: 27 Message-ID: <368118f0.12180594@news.uunet.be> References: <755ubm$quj$1@irk.zetnet.co.uk> <2uFd2.2296$WZ6.6110370@ptah.visi.com> NNTP-Posting-Host: pool02b-194-7-148-23.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newscore.univie.ac.at!newsfeed.ecrc.net!newsfeed.icl.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach seebs@plethora.net (Peter Seebach) on Wed, 16 Dec 1998 03:07:42 GMT to alt.folklore.computers: > In article , > Marco S Hyman wrote: > >Snicker! As someone once said "never underestimate the power of human > >stupidity!" Just because a system call or library function exists > >does NOT mean it will be used. Or have you never seen this code (or > >something very like it) in random applications: > > > for ( i = 0; i < count; i++ ) { > > dst[ i ] = src[ i ]; > > } > > Not only that, but once I benchmarked a test program, and memcpy was 280 > times faster on the Sun workstation involved. Not sure why. :) _Two_hundred_and_eighty_ times? Must have been that the C code was interrupted (what else was running?), and memcpy wasn't (or was less). This also reminded me of a PC speedup utility someone wrote about in this group, that made the clock run slower... -- "Managing senior programmers is like herding cats." --Dave Platt ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 21:22:32 GMT Organization: University of Illinois at Urbana-Champaign Lines: 26 Message-ID: <7598ao$p89$1@vixen.cso.uiuc.edu> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!128.174.5.49!vixen.cso.uiuc.edu!tsmurphy In article <368118f0.12180594@news.uunet.be>, Luc Van der Veken wrote: >Also sprach seebs@plethora.net (Peter Seebach) on Wed, 16 Dec >1998 03:07:42 GMT to alt.folklore.computers: >> > for ( i = 0; i < count; i++ ) { >> > dst[ i ] = src[ i ]; >> > } >> >> Not only that, but once I benchmarked a test program, and memcpy was 280 >> times faster on the Sun workstation involved. Not sure why. :) > >_Two_hundred_and_eighty_ times? >Must have been that the C code was interrupted (what else was >running?), and memcpy wasn't (or was less). It's not too unbelievable and certainly a machine with a poor branch prediction algorithm, a high wrong branch taken penalty, and a penalty for unaligned accesses would come close to this. I wrote a self-modifying version of strcpy (using descriptor based strings) on Linux which was something like 20x faster than the regular Linux strcpy (which uses the above, brain-dead algorithm) -- Terry ###### Newsgroups: alt.folklore.computers From: alderson@netcom.netcom.com (Richard M. Alderson III) Subject: Re: Is VMS taboo, or what? In-Reply-To: martin@ibert.com's message of Tue, 15 Dec 1998 20:52:34 GMT Message-ID: Sender: alderson@netcom.netcom.com Reply-To: alderson@netcom.com Organization: NETCOM On-line services References: <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <74j13q$6c3$1@news.wlg.netlink.net.nz> <1djrsod.6w627whz8fvgN@[10.0.0.3]> <36775424.60404627@news.uunet.be> <36728afd.10991534@news.uunet.be> <3681cb9a.15103630@news.home.ibert.com> Date: Wed, 16 Dec 1998 22:44:53 GMT Lines: 31 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!netnews.com!ix.netcom.com!netcom!alderson In article <3681cb9a.15103630@news.home.ibert.com> martin@ibert.com (Martin Ibert) writes: >On Tue, 15 Dec 1998 00:19:50 GMT, alderson@netcom.netcom.com >(Richard M. Alderson III) wrote: >>That's an awful lot of trouble to go to just to save a minuscule amount of >>disk space, especially since the OS isn't going to let you umount /usr and >>/etc while you're running. The only way what was originally suggested makes >>any sense at all is if you *could* write to unmounted filesystems. Anything >>else is just schratzing around for the sake of schratzing around. >I can see another reason why you'd like to do that: to be able to mount / >read-only. You cannot expect to run in production with all files in /etc >read-only, but having a read-only root in production has one hell of a lot of >reasons going for it, so having /etc on a separate filesystem makes a _lot_ of >sense to me. No, that's not what seems like a lot of trouble to me. The previous discussion was of leaving a few files in the underlying directories which serve as mount points (specifically, in my example, /usr). The post to which you responded was my own response to someone who suggested *copying* the files from the underlying directory to the mounted filesystem on boot, and back again on shut- down. Now, I don't know about you, but to me that seems like a lot to bother with for not much gain (if any). -- Rich Alderson Last LOTS Tops-20 Systems Programmer, 1984-1991 Current maintainer, MIT TECO EMACS (v. 170) last name @ XKL dot COM Chief systems administrator, XKL LLC, 1998-now ###### From: David Scheidt Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 16 Dec 1998 23:18:16 GMT Organization: EnterAct L.L.C. Turbo-Elite News Server Lines: 55 Message-ID: <759f3o$oq5$1@eve.enteract.com> References: <1456.648T1620T6105137@sky.bus.com> <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <367af967.61373440@news.uunet.be> <757341$m8$1@eve.enteract.com> <36831906.12202596@news.uunet.be> NNTP-Posting-Host: 207.229.143.6 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.enteract.com!news.enteract.com!dscheidt Luc Van der Veken wrote: : Also sprach David Scheidt on 16 Dec 1998 : 01:41:21 GMT to alt.folklore.computers: : > No. It doesn't. The Windows box I have on my desk at work, which is there : > only so I can read my email is not easy to use. It simply doesn't do what : > I want it to. : I suppose you mean it doesn't interprete your commands the way : you are accustomed to. Comfort: true the other way around too. No. I meant "It doesn't do what I want it to do, because the software won't let it." : Awk for example is typical for unix (at least I think it is) - if : you want to do the same thing under windows, the way MS would : probably like you most to do it is to open the message in Word : and write a Word macro to do the work. This assumes the existence of MS word on the machine in question. It ain't there, and I can't get it without spending money. Since the *only* reason it runs windows is I don't know of a third party product that talks to an exchange server, I am hard pressed to justify spending money. Awk, on the other hand, has been on almost every Un*x box I have worked on, if only because I built and installed it, as there are versions to be had for free. : The awk script would be shorter, but then again if you wanted to : do something entirely different you'd have to switch to another : scripting language. Under windows you stay with the same: the : [dis]advantages of general purpose languages. Awk does a small number of things very well. If I want to do something else, I use a tool designed to do what I want. If one doesn't exist, it will pretty quickly. : URL's under the letter U), click edit. : Probably all it takes is setting another action as the default. : If not, select the type of action you want to change (you'll want : the one in bold print, = default, now) and click edit again, or : create an all new action. Enter the command line you want the : system to use. BTDT. Everything points towards the internet exploder. : Or, you can solve it like most users do: reformat your harddisk, : and reinstall Windows. Last time I reinstalled Unix because of application failure: never. David -- David Scheidt Large fibreglass fruits are much the same the world over. -- Vicki Parslow Stafford ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 09:32:39 GMT Organization: Daedalus Consulting Lines: 63 Message-ID: <75aj3n$9hu$1@news.wlg.netlink.net.nz> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Luc Van der Veken wrote: >Also sprach seebs@plethora.net (Peter Seebach) on Wed, 16 Dec >1998 03:07:42 GMT to alt.folklore.computers: >> > for ( i = 0; i < count; i++ ) { >> > dst[ i ] = src[ i ]; >> > } >> >> Not only that, but once I benchmarked a test program, and memcpy was 280 >> times faster on the Sun workstation involved. Not sure why. :) > >_Two_hundred_and_eighty_ times? >Must have been that the C code was interrupted (what else was >running?), and memcpy wasn't (or was less). I haven't had anything to do with the Sparc at assembly level, but I'm not all that surprised. RISC architectures are generally not good at byte ops -- they're much happier shuffling data around in their natural word sizes. So memcpy() (and remember that memcpy knows the length in advance, so can do a bit of preparatory work before starting the bulk of the copy) will in most cases[1] be copying in 32 bit chunks to longword aligned locations, almost certainly in an unrolled loop. On the other hand, the loop above is copying byte chunks. For every single 32 bit transfer memcpy() does, the above loop will go around four times, *including* the boundary tests. Unroll the loop four times, and memcpy does four longword copies plus a loop test and a branch, while the loop above does sixteen byte copies, sixteen tests and sixteen branches. Then there's the copies themselves. To do an unaligned byte transfer on a 32 bit RISC machine you have to shift and mask. Worse, to store, it has to read the 32 bit destination and shift/mask in the byte. So the loop, all up, does for each 4 bytes four 32 bit loads from source to register three shifts and four masks to register four 32 bit reads from destination to register three shifts and four masks to register four 32 bit stores four increments four tests four branches Oh yes, and branches are expensive because they dump the pipeline. On the other hand, a deftly coded [1] memcpy, does one 32 bit copy, and since it'll be all nicely unrolled, it won't branch until its done a bunch of them. [1] Things are fast if both source and destination have the same alignment. Pointers returned from malloc() or naturally aligned in strutures will be like that, so it happens more often than not, but memcpy() will degrade if both strings are not the same distance away from the natural word size. Note that (say) src=0x1001 to dest=0x2001 can still be fast -- you have to special-case the longword transfer from 0x1000 to 0x2000 to not stomp on the byte at 0x2000, but after that everything is simple 32 bit transfers (until the end when you'll inevitably have to juggle things again). It's src=0x1001 to dest=0x1002 where things come unstuck. But it can still be coded faster than the ignorant byte copy. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 09:46:21 -0000 Lines: 33 Message-ID: <3678d2ef.0@192.168.0.20> References: <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <36767e0a.0@192.168.0.20> <1998Dec15.190813.10619@picarefy.picarefy.com> <36779525.0@192.168.0.20> <3679edd7.1148000@news.uunet.be> NNTP-Posting-Host: 195.147.136.27 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 Luc Van der Veken wrote in message <3679edd7.1148000@news.uunet.be>... >Also sprach "Samael" on Wed, 16 Dec 1998 >11:10:04 -0000 to alt.folklore.computers: > >> >> James W. Birdsall wrote in message >> >This is why so many people want keyboard shortcuts. Menus *are* handy for >> >infrequently-used functions, but being forced to go through them just slows >> >down experienced users. >> >> Certianly do. Which is why I use the keyboard for most things. > > Me too >But depending somewhat on what program it is: some are less >keyboard friendly than they ought to be. >[AAMOF, there may be a key combination for it, but this afternoon >I grabbed my mouse to center a paragraph in Word because ctrl/c >didn't do the job] Ctrl-C is the windows standard for 'copy' as in 'copy and paste'. Word has therfore moved to using the next available character for 'center', in this case 'e'. Ctrl-E is therefore used. Annoying , I know, but as Ctrl-C is already used by windows, not a lot they could do. Samael ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 16:24:08 GMT Organization: University of Illinois at Urbana-Champaign Lines: 24 Message-ID: <75bb78$n29$1@vixen.cso.uiuc.edu> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer1.sprintlink.net!news.sprintlink.net!vixen.cso.uiuc.edu!tsmurphy In article <75aj3n$9hu$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >[1] Things are fast if both source and destination have the same >alignment. Pointers returned from malloc() or naturally aligned in >strutures will be like that, so it happens more often than not, but >memcpy() will degrade if both strings are not the same distance away >from the natural word size. Note that (say) src=0x1001 to dest=0x2001 >can still be fast -- you have to special-case the longword transfer from >0x1000 to 0x2000 to not stomp on the byte at 0x2000, but after that >everything is simple 32 bit transfers (until the end when you'll >inevitably have to juggle things again). It's src=0x1001 to dest=0x1002 >where things come unstuck. But it can still be coded faster than the >ignorant byte copy. How precisely would you do the src=0x1001 to dest=0x1002 faster than using a byte copy? The only way I can think of it to still only do dword accesses (after you reach the right point) and cache the data in the registers, and shift it into complete dwords, and then write it back. But it seems that this would barely be faster than simply doing byte access, since you'd have to do an extra two or three instructions in the loop. (though I haven't benchmarked it yet). -- Terry ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 18:56:40 GMT Organization: . Lines: 20 Message-ID: <367f4ae0.2294068@news.uunet.be> References: <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <36767e0a.0@192.168.0.20> <1998Dec15.190813.10619@picarefy.picarefy.com> <36779525.0@192.168.0.20> <3679edd7.1148000@news.uunet.be> <3678d2ef.0@192.168.0.20> NNTP-Posting-Host: pool02b-194-7-84-215.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!news-lond.gip.net!news.gsl.net!gip.net!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach "Samael" on Thu, 17 Dec 1998 09:46:21 -0000 to alt.folklore.computers: > Ctrl-C is the windows standard for 'copy' as in 'copy and paste'. Word has > therfore moved to using the next available character for 'center', in this > case 'e'. Ctrl-E is therefore used. > > Annoying , I know, but as Ctrl-C is already used by windows, not a lot they > could do. Now that you say it, I use ctrl-c to copy all the time (in developer studio etc.), but didn't think of that. Just "bold is ctrl-b, italic is ctrl-i, so center must be ctrl-c" -- News headlines. Earlier today, Saddam appealed onto all his countrymen to start a Jihad against men who cheat on their wives. ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 18:56:41 GMT Organization: . Lines: 17 Message-ID: <36804ae4.2298605@news.uunet.be> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <7598ao$p89$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: pool02b-194-7-84-215.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!rill.news.pipex.net!pipex!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach tsmurphy@cs.uiuc.edu (TERENCE MURPHY) on 16 Dec 1998 21:22:32 GMT to alt.folklore.computers: > I wrote a self-modifying version of strcpy (using descriptor based > strings) on Linux which was something like 20x faster than the regular > Linux strcpy (which uses the above, brain-dead algorithm) Still having problems believing 280x slowdown just because of doing it "by hand" in C code instead of using a library routine. 20x, or even 50x: OK, but what we have here is 14 times 20x. Even if you drop all pipelining etc., it still is a lot. -- News headlines. Earlier today, Saddam appealed onto all his countrymen to start a Jihad against men who cheat on their wives. ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 18:56:42 GMT Organization: . Lines: 10 Message-ID: <36874dd4.3050706@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> NNTP-Posting-Host: pool02b-194-7-84-215.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail 426 messages in 24 days, at least 2 OS wars in subthreads, yet nobody said anything about starting a new group so far. How many more do we need before it's created automatically? -- News headlines. Earlier today, Saddam appealed onto all his countrymen to start a Jihad against men who cheat on their wives. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 08:28:15 +0000 Organization: Tanglewood Message-ID: <3678C09F.D1F10D66@tnglwood.demon.co.uk> References: <367541b4.0@192.168.0.20> <753imq$1jg$1@agate.berkeley.edu> <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913922896 nnrp-07:25569 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 23 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!diablo.theplanet.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Mark Wilkins wrote: > state of "knowing what you're doing." Though these artists may not have > the same strengths in symbolic manipulation that I do, they are able to > use their spatial memory to be productive in ways I could never dream > of. I have spent a lot of my life making broadcasting equipment, and I can confirm what you say. However, with a lot of our control panels, we have an essentially "key per function" approach, with keys either engraved, or labelled on displays. We find that expert users tend to look away from both our displays and the keyboard, fix their eyes on the picture monitor, and "play by ear" on the keyboard. The trick to doing this is to remember, spatially, that "position" is up there, "size" is down here, "rotation" is along to the side, and so on. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 08:39:28 +0000 Organization: Tanglewood Message-ID: <3678C340.B9D1BD33@tnglwood.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 913922897 nnrp-07:25569 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 17 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.wirehub.nl!diablo.theplanet.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Martin Ibert wrote: > - The most feauritis-infested, bulky, unstable, unpleasant, klugy > development environment ever. You don't like UN*X source-code > debuggers? Try Microsoft's. You'll love gdb after that. (At least > Microsoft Visual Nightmare doesn't take all night to crash.) Even better, use Quick C for 3.1, and set a breakpoint in the handler for a system modal dialogue box. The resulting disintegration of the entire windowing system is very pretty, and takes several minutes to play out to completion. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 19:26:33 GMT Organization: Laser-Scan Ltd. Lines: 34 Message-ID: <75blt9$dlq@relay.lsl.co.uk> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <7598ao$p89$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!news.maxwell.syr.edu!rill.news.pipex.net!pipex!join.news.pipex.net!pipex!news.lsl.co.uk!benh TERENCE MURPHY (tsmurphy@cs.uiuc.edu) wrote: : In article <368118f0.12180594@news.uunet.be>, : Luc Van der Veken wrote: : >Also sprach seebs@plethora.net (Peter Seebach) on Wed, 16 Dec : >1998 03:07:42 GMT to alt.folklore.computers: : >> > for ( i = 0; i < count; i++ ) { : >> > dst[ i ] = src[ i ]; : >> > } : >> : >> Not only that, but once I benchmarked a test program, and memcpy was 280 : >> times faster on the Sun workstation involved. Not sure why. :) : > : >_Two_hundred_and_eighty_ times? : >Must have been that the C code was interrupted (what else was : >running?), and memcpy wasn't (or was less). : It's not too unbelievable and certainly a machine with a poor branch : prediction algorithm, a high wrong branch taken penalty, and a penalty : for unaligned accesses would come close to this. So you think memcpy doesn't use a branch, and the compiler misaligned those arrays? ;-) : I wrote a self-modifying version of strcpy (using descriptor based : strings) on Linux which was something like 20x faster than the regular : Linux strcpy (which uses the above, brain-dead algorithm) The relative speed will surely be highly dependent on the length of strings. The speed penalty for modifying code can be very high on recent processors (which use separate caches). -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 17 Message-ID: Date: Thu, 17 Dec 1998 19:34:49 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913923289 205.166.146.8 (Thu, 17 Dec 1998 13:34:49 CDT) NNTP-Posting-Date: Thu, 17 Dec 1998 13:34:49 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <75aj3n$9hu$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >I haven't had anything to do with the Sparc at assembly level, but I'm >not all that surprised. This wasn't a Sparc; it was a 3/60. The content was a loop of copying 1,000,000 bytes of data from one array to another and then back. It wasn't fast enough that I would think it had optimized everything away. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 19:36:24 GMT Organization: Daedalus Consulting Lines: 35 Message-ID: <75bmfo$59g$1@news.wlg.netlink.net.nz> References: <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> <75bb78$n29$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <75bb78$n29$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: >How precisely would you do the src=0x1001 to dest=0x1002 faster than >using a byte copy? The only way I can think of it to still only do >dword accesses (after you reach the right point) and cache the data >in the registers, and shift it into complete dwords, and then write >it back. But it seems that this would barely be faster than simply >doing byte access, since you'd have to do an extra two or three >instructions in the loop. (though I haven't benchmarked it yet). What I had in mind was that for the case where the alignement was one off, you'd go: fetch 32 bit source to reg 1 mask bottom byte to reg 2 shift reg 1 left 8 bits shift in bottom byte from reg 3 to top store 32 bits The next pass (this would be unrolled) would use reg 3 for reg 2 and vice-versa. This has the beauty that it only does 32 bit transfers to memory -- there's a bit of jiggling going on in registers, but that's comparitively fast. Just unrolling the loop is quite a big win on RISC boxes, although if you do it too much you start giving the caches a hard time. Key thing is to put unrolled loops in hot bits of code, memcpy() & memcmp() for eg... (I could start a tirade here about why null-terminated strings are Bad Things on RISC boxes, but if you've been paying attention you'll have worked that out already. 8-) -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: "Donald Tees" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 20:09:37 -0500 Organization: IGS - Information Gateway Services Lines: 42 Message-ID: <75c9cp$9l7$1@news.igs.net> References: <367541b4.0@192.168.0.20> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <1998Dec17.205055.27024@lorelei.approve.se> NNTP-Posting-Host: ttye08.kw.igs.net X-Trace: news.igs.net 913942745 9895 206.248.37.136 (18 Dec 1998 00:59:05 GMT) X-Complaints-To: abuse@igs.net NNTP-Posting-Date: 18 Dec 1998 00:59:05 GMT X-Newsreader: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!cyclone.bc.net!torn!nntp.igs.net!news.igs.net!not-for-mail I think you are missing a central point. To talk about pictures with words is as silly as talking about words using pictures. I have been programming for thirty years now, and I also miss a command line interface. I can do (express to the computer) things with commands that I cannot do with a mouse. However, to extrapolate from that that anyone preferring a mouse is too stupid to write their name is a rather big jump. I have difficulty imagining a good command line interface for PC-Paint. There are tasks that are suited to mouses. There are tasks that are suited to words. There are minds that work best vocally and minds that work best visually. The real problem is the provincialism of the programmer that claims *their* way is the *only* way, coupled with OS styles that fall prey to that narrow-mindedness. Goran Larsson wrote in message <1998Dec17.205055.27024@lorelei.approve.se>... >In article <3676F14B.4E02ABDF@the_bottom_for_my.e-mail>, >Mark Wilkins wrote: > >> ... Many of the artists require repeated explanations to >> remember a textual command but only need to be shown a menu item once. > >The way you describe these artists makes me wonder if they are >able to use a pen to type their own name on a piece of paper. > >Why is it so much harder to remember words and how to type them >if the receiver of the typed message is a computer? > >Perhaps we are closing the circle and will revert back to using >hieroglyphs instead of the alphabet, one of mankinds greatest >inventions, just because some artists can't be bothered to learn >how to type simple words. > >-- > Goran Larsson hoh AT approve DOT se > I was an atheist, http://home DOT swipnet DOT se/hoh/ > until I found out I was God. ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec17.205055.27024@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <367541b4.0@192.168.0.20> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> Lines: 21 Date: Thu, 17 Dec 1998 20:50:55 GMT NNTP-Posting-Host: 130.244.179.46 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 913928318 130.244.179.46 (Thu, 17 Dec 1998 21:58:38 MET DST) NNTP-Posting-Date: Thu, 17 Dec 1998 21:58:38 MET DST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article <3676F14B.4E02ABDF@the_bottom_for_my.e-mail>, Mark Wilkins wrote: > ... Many of the artists require repeated explanations to > remember a textual command but only need to be shown a menu item once. The way you describe these artists makes me wonder if they are able to use a pen to type their own name on a piece of paper. Why is it so much harder to remember words and how to type them if the receiver of the typed message is a computer? Perhaps we are closing the circle and will revert back to using hieroglyphs instead of the alphabet, one of mankinds greatest inventions, just because some artists can't be bothered to learn how to type simple words. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: "T.W. Seddon" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 21:30:19 GMT Organization: University of Newcastle upon Tyne Lines: 19 Message-ID: <75bt5b$so5$1@ucsnew1.ncl.ac.uk> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <7598ao$p89$1@vixen.cso.uiuc.edu> <36804ae4.2298605@news.uunet.be> NNTP-Posting-Host: aidan.ncl.ac.uk X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news-stock.gip.net!news-peer.gip.net!news-lond.gip.net!news.gsl.net!gip.net!nntp.news.xara.net!xara.net!server6.netnews.ja.net!news.ncl.ac.uk!aidan!n5013784 Luc Van der Veken (lucvdv@null.net) wrote: > Still having problems believing 280x slowdown just because of > doing it "by hand" in C code instead of using a library routine. > 20x, or even 50x: OK, but what we have here is 14 times 20x. Even > if you drop all pipelining etc., it still is a lot. The C library routine may use an instruction specially designed for moving bytes (eg movsl (stosl? I forget) on the 386) which can often turn out far quicker than individual byte moves. I don't know whether the suns have such an instruction though. Apropos of nothing much is there any guarantee about what will happen using memcpy() vs a for loop if the array being copied is declared as volatile? -- -- --Tom I do not have time to brainwash you, so listen carefully ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 21:36:25 GMT Organization: University of Illinois at Urbana-Champaign Lines: 20 Message-ID: <75btgp$b6i$1@vixen.cso.uiuc.edu> References: <75aj3n$9hu$1@news.wlg.netlink.net.nz> <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <75bmfo$59g$1@news.wlg.netlink.net.nz>, Don Stokes wrote: >(I could start a tirade here about why null-terminated strings are Bad >Things on RISC boxes, but if you've been paying attention you'll have >worked that out already. 8-) Oh, absolutely. The descriptor routines which I wrote and alluded to earlier which made a strcpy() which was 20x faster than stock Linux strcpy() used thse techniques we're talking about -- and it was on a P5 which probably doesn't even have the byte access penalties that the P6 does (though I don't know P5 architecture very well). Implemented on an Alpha (which is my next project), I bet the peformance would be much more compelling. I understand than Alpha does byte accesses (at least up to EV5) very poorly and that's the reason for poor Windows performance (surely Unix is affected the same way, since it is written in C?) -- Terry ###### From: Mike Swaim Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <753tse$91s$1@agate.berkeley.edu> <3676083D.E3938AB2@tnglwood.demon.co.uk> <755h9n$deh$1@ligarius.ultra.net> <1704.653T2288T6175773@sky.bus.com> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <2230.653T2739T10516280@sky.bus.com> Organization: PointeCom User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/2.2.8-RELEASE (i386)) Lines: 15 Message-ID: NNTP-Posting-Date: Thu, 17 Dec 1998 15:41:38 CDT X-Trace: sv1-avMV3yym+Ajx3NfgHjD3X0ySjzK5TNQln05E9n6LprdMKqp0aXRyt6e4hr58ihs6Frjn2QtiY9CE9IV!LtAftzMqyuQ= X-Complaints-To: abuse@GigaNews.Com X-Abuse-Info: Please be sure to forward a copy of ALL headers X-Abuse-Info: Otherwise we will be unable to process your complaint properly Date: Thu, 17 Dec 1998 21:41:38 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!newshub.northeast.verio.net!newsfeed.cwix.com!207.207.0.26!nntp.giganews.com!news1.giganews.com.POSTED!not-for-mail Charlie Gibbs wrote: : A fascinating alternate view. Thanks for pointing that out. : You've illustrated two totally different forms of abstraction. : And I'd say that's one of the most convincing arguments yet : that a system should provide both textual and graphic interfaces. AutoCad does that. The last place I worked at had a number of dual headed machines (one graphic monitor, one text monitor) for the CAD people. I suppose that with multimonitor support in Windows now, they could do it again. -- Mike Swaim, Avatar of Chaos: Disclaimer:I sometimes lie. Home: swaim@c-com.net Alum: swaim@alumni.rice.edu Quote: "Boingie"^4 Y,W&D ###### From: dpeschel@u.washington.edu (D. Peschel) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 22:11:06 GMT Organization: University of Washington, Seattle Lines: 25 Message-ID: <75bvhq$evg$1@nntp1.u.washington.edu> References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <2230.653T2739T10516280@sky.bus.com> NNTP-Posting-Host: saul5.u.washington.edu X-Trace: nntp1.u.washington.edu 913932666 15344 (None) 140.142.17.35 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: dpeschel Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!news.u.washington.edu!dpeschel In article , Mike Swaim wrote: >Charlie Gibbs wrote: >: A fascinating alternate view. Thanks for pointing that out. >: You've illustrated two totally different forms of abstraction. >: And I'd say that's one of the most convincing arguments yet >: that a system should provide both textual and graphic interfaces. I don't know if Charlie meant, "provide them on two separate monitors at the same time"! But I'm sure it's a useful feature, especially for debugging, though I've never used it. > AutoCad does that. The last place I worked at had a number of dual >headed machines (one graphic monitor, one text monitor) for the CAD >people. I suppose that with multimonitor support in Windows now, they >could do it again. Except they wouldn't do it as well as before. Windows' text support is pretty uneven; I don't think it supports the standard features of VGA or even EGA cards (viz., 64 colors of text, user-defined character sets, 512-byte character sets, underlining). -- Derek ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <7598ao$p89$1@vixen.cso.uiuc.edu> <36804ae4.2298605@news.uunet.be> <75bt5b$so5$1@ucsnew1.ncl.ac.uk> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 25 Message-ID: Date: Thu, 17 Dec 1998 22:15:34 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 913932934 205.166.146.8 (Thu, 17 Dec 1998 16:15:34 CDT) NNTP-Posting-Date: Thu, 17 Dec 1998 16:15:34 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <75bt5b$so5$1@ucsnew1.ncl.ac.uk>, T.W. Seddon wrote: >Apropos of nothing much is there any guarantee about what will happen >using memcpy() vs a for loop if the array being copied is declared as >volatile? I see nothing special about it; I think C9X adds a sequence point after the function call, just in case, but I could be confused. I wouldn't want to rely on the contents of the target before the next sequence point. Hmm. Y'know, you have an interesting point there. My thinking is that there's no guarantee of anything special, because there's no indication of how many accesses are involved in memcpy - but of course, with the right sorts of volatile data, you would care. Hmm. I think I'll ask about that in comp.std.c. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 22:40:01 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <75c3g5$cbd$3@roch.zetnet.co.uk> NNTP-Posting-Host: d1.56.d2.d8 X-Server-Date: 18 Dec 1998 04:40:03 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <75c3g5$cbd$3@roch.zetnet.co.uk>, lisard@zetnet.co.uk wrote: >It's logical, but strange, to hear of accountants being busy with >year-end stuff in December, rather than March. Yeah, my client's financial year is the calendar year. I just do what they tells me. -- Howard S Shubs The Denim Adept ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 17 Dec 1998 22:40:01 -0600 Organization: MindSpring Enterprises Lines: 9 Message-ID: References: <75c3g5$cbd$3@roch.zetnet.co.uk> NNTP-Posting-Host: d1.56.d2.d8 X-Server-Date: 18 Dec 1998 04:40:03 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.cwix.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <75c3g5$cbd$3@roch.zetnet.co.uk>, lisard@zetnet.co.uk wrote: >It's logical, but strange, to hear of accountants being busy with >year-end stuff in December, rather than March. Yeah, my client's financial year is the calendar year. I just do what they tells me. -- Howard S Shubs The Denim Adept ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 23:18:29 GMT Message-ID: <75c3g5$cbd$3@roch.zetnet.co.uk> References: NNTP-Posting-Host: man-051.dialup.zetnet.co.uk X-Trace: roch.zetnet.co.uk 913936709 12653 194.247.41.63 (17 Dec 1998 23:18:29 GMT) NNTP-Posting-Date: 17 Dec 1998 23:18:29 GMT X-Everything: Net-Tamer V 1.08X Lines: 12 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!newsfeed1.swip.net!swipnet!rill.news.pipex.net!pipex!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-15 hshubs@mindspring.com(HowardSShubs) said: :I explained the symbolic method to my troubled user today, and I :think he got it! The accountant may have understood the :binary/octal method, but I've not quizzed her yet. She's kinda :busy with year-end stuff. It's logical, but strange, to hear of accountants being busy with year-end stuff in December, rather than March. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 23:18:31 GMT Lines: 31 Message-ID: <75c3g7$cbd$4@roch.zetnet.co.uk> References: <1704.653T2288T6175773@sky.bus.com> NNTP-Posting-Host: man-051.dialup.zetnet.co.uk X-Trace: roch.zetnet.co.uk 913936711 12653 194.247.41.63 (17 Dec 1998 23:18:31 GMT) NNTP-Posting-Date: 17 Dec 1998 23:18:31 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!195.200.0.51.MISMATCH!newshub.bart.net!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-15 cgibbs@sky.bus.com said: :The paradox arises because most people don't realize that "easy to :learn" is not the same as "easy to use". Menus are great if you're :noodling around trying to get the lay of the land, but they just get :in the way once you know what you're doing. Of course, nowadays :your average luser never gets to know what he's doing, either :because he doesn't use a program frequently enough to learn it, :or because he refuses to learn, or because he quickly throws the :program away in favour of next month's fad. Thus the myth that :command lines are "difficult to use" gets perpetuated. Or because there just isn't another way through than to fight with the menus. Most people know what they need to get their jobs done and don't bother with the rest, I suspect. Unfortunately, that means they never get as far as the key redefinition stuff in Word, and even if they did, which keystroke set would they switch to? Anything else would require a new learning curve; so there's no realistic option but to stick with the confusing, cluttered, condescending commandset they've already learned. Whereas I've been using Unix, even with the CDE desktop manager, for the last few weeks at work, and have not touched a menu in... oh, ages. (Last time I had to set up the right font for xterm. Choosing 'large' is easier than remembering 80 characters of exact font specifier.) Except for testing the app I'm writing, of course, but since I'm not using that as such it doesn't count. NOBODY is taking my AlphaStation away from me! NOBODY!!!!! };> -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 23:18:33 GMT Message-ID: <75c3g9$cbd$5@roch.zetnet.co.uk> References: <36780619.96636558@Rockyd> NNTP-Posting-Host: man-051.dialup.zetnet.co.uk X-Trace: roch.zetnet.co.uk 913936713 12653 194.247.41.63 (17 Dec 1998 23:18:33 GMT) NNTP-Posting-Date: 17 Dec 1998 23:18:33 GMT X-Everything: Net-Tamer V 1.08X Lines: 13 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-16 alex*@*rockvax.rockefeller.edu(AlexandrePechtchanski) said: :>>And while we are at it, how in "hierarchical" Unix file system :>>one can determine the (used) size of "/" partition (excluding :other file systems hanging off /) ? > [not vstatfs] df ? Unless I've misunderstood the question, it'll show you every filesystem with the mount point and the space eaten/remaining/total. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 17 Dec 1998 23:50:45 GMT Organization: Daedalus Consulting Lines: 37 Message-ID: <75c5cl$cg7$1@news.wlg.netlink.net.nz> References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don In article <75btgp$b6i$1@vixen.cso.uiuc.edu>, TERENCE MURPHY wrote: >Oh, absolutely. The descriptor routines which I wrote and alluded to >earlier which made a strcpy() which was 20x faster than stock Linux >strcpy() used thse techniques we're talking about -- and it was on a P5 >which probably doesn't even have the byte access penalties that the P6 >does (though I don't know P5 architecture very well). *Any* 32 bit architecture will be penalised by having to do four byte-wide transfers instead of one longword transfer, if only due to the higher instruction count. >Implemented on an Alpha (which is my next project), I bet the >peformance would be much more compelling. I understand than Alpha does >byte accesses (at least up to EV5) very poorly and that's the reason >for poor Windows performance (surely Unix is affected the same way, >since it is written in C?) C itself is perfectly capable of dealling with lengths and pointers. Where it falls down is the abortion that is the standard library, which in turn influence and are influenced by by Unix parameter passing conventions, and lack of ability to deal with length/pointer pairs as single objects. Basically, C really just gives you the machine objects and lets you have at it; it's no better at handling descriptors as null-terminated strings (apart from the fact that "string" is null-terminated). Back when I was doing a lot of C under VMS, doing a lot of VMSsy things, I'd quite often tell the C library to go take a running jump, and use VMS's libraries of descriptor based string manipulation routines. It was actually easier that way, since all the VMS stuff I was calling expected VMS-style descriptors. But when you're manipulating stuff to be used with Unix system & library calls, you're largely forced to use null-terminated strings. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 04:15:16 +0100 Organization: [Posted via] Interactive Networx Lines: 12 Message-ID: <1dk6q0n.1oe4e2ldrk1uaN@n37-5.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> NNTP-Posting-Host: n37-5.berlin.snafu.de User-Agent: MacSOUP/2.4b4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!newsfeed.wirehub.nl!newscore.ipf.de!unlisys!news.snafu.de!jnickelsen Howard S Shubs wrote: > Success is always an odd value, with the best success code being STATUS % > 8 == 1. I wish I had known this when I ported a "make" program to VMS. Being quite unexperienced then, I did not know where to look and figured out the common success codes of the compiler and linker by trying. At least that didn't take too long. -- Juergen Nickelsen ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 04:15:19 +0100 Organization: [Posted via] Interactive Networx Lines: 11 Message-ID: <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> NNTP-Posting-Host: n37-5.berlin.snafu.de User-Agent: MacSOUP/2.4b4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!wuff.mayn.de!unlisys!news.snafu.de!jnickelsen phil wrote: > A GUI is a bitch to develop for. Of all those available, NT is probably > the easiest (you can produce a console app with minimal grief). I agree that console applications under Windows NT are fairly straightforward (and can even be ported from Unix with little effort). But this is not in any way related to *GUI* applications. -- Juergen Nickelsen ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 04:41:20 +0100 Organization: [Posted via] Interactive Networx Lines: 11 Message-ID: <1dk6q7m.fap2ctifdx6oN@n65-158.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> NNTP-Posting-Host: n65-158.berlin.snafu.de User-Agent: MacSOUP/2.4b4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!212.63.192.161.MISMATCH!newshub.bart.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.tli.de!unlisys!news.snafu.de!jnickelsen phil wrote: > A GUI is a bitch to develop for. Of all those available, NT is probably > the easiest (you can produce a console app with minimal grief). I agree that console applications under Windows NT are fairly straightforward (and can even be ported from Unix with little effort). But this is not in any way related to *GUI* applications. -- Juergen Nickelsen ###### From: fleetvis@ricochet.net (phil) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 06:41:36 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 20 Message-ID: <3679f5fb.9052847@news.ricochet.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> NNTP-Posting-Host: 204.179.136.245 X-Trace: 913962973 F7W8ZAFJM88F5CCB3C usenet76.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newspump.monmouth.com!newspeer.monmouth.com!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail On Fri, 18 Dec 1998 04:15:19 +0100, jnickelsen@acm.org (Juergen Nickelsen) wrote: >phil wrote: > >> A GUI is a bitch to develop for. Of all those available, NT is probably >> the easiest (you can produce a console app with minimal grief). > >I agree that console applications under Windows NT are fairly >straightforward (and can even be ported from Unix with little effort). >But this is not in any way related to *GUI* applications. > Except that i thought that the argument was that NT was difficult to develop for, and i believe that it is not NT that is messy to program, but rather the Windows GUI. It's not so bad these days, but i'm still surprised about the amount of infrastructure that has to be put into place for even the simplest Windows app. phil. ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 09:50:34 -0000 Lines: 15 Message-ID: <367a265f.0@192.168.0.20> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n65-158.berlin.snafu.de> NNTP-Posting-Host: 195.147.134.81 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 >phil wrote: > >> A GUI is a bitch to develop for. Of all those available, NT is probably >> the easiest (you can produce a console app with minimal grief). Why is a GUI a bitch to develop for? I have programmed for GUI's and straight text app's and I find them equally easy, just different. Samael ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 98 12:50:33 GMT Organization: UltraNet Communications, Inc. Lines: 24 Message-ID: <75dkgn$55a$2@ligarius.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <36874dd4.3050706@news.uunet.be> NNTP-Posting-Host: d7.dial-12.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 18 Dec 1998 13:15:03 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!feed1.news.rcn.net!rcn!news.ultranet.com!d7 In article <36874dd4.3050706@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >426 messages in 24 days, at least 2 OS wars in subthreads, yet >nobody said anything about starting a new group so far. But the content has been very good for newbies (and even us oldbies). The thread has brought up some very good points to consider when implementing software both at the OS level and the application level. For those who have been reading the stuff objectively, it should have been clear that there isn't (and never can be) _One True Program_. Up until now, I'd been very discouraged about people here being able to fight w/o resorting to religiousity. This thread and a few other changed that :-). > >How many more do we need before it's created automatically? > 377777777777-^D426 /BAH Subtract a hundred and four for e-mail. ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 18 Dec 1998 16:04:34 GMT Organization: University of Illinois at Urbana-Champaign Lines: 18 Message-ID: <75duei$70a$1@vixen.cso.uiuc.edu> References: <73ela0$51h@hermes.acs.unt.edu> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> <3679f5fb.9052847@news.ricochet.net> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!howland.erols.net!vixen.cso.uiuc.edu!tsmurphy In article <3679f5fb.9052847@news.ricochet.net>, phil wrote: >Except that i thought that the argument was that NT was difficult to >develop for, and i believe that it is not NT that is messy to program, >but rather the Windows GUI. It's not so bad these days, but i'm still >surprised about the amount of infrastructure that has to be put into >place for even the simplest Windows app. Is there any GUI environment where you -don't- have to do this? A decent "hello world" program in Xlib is a couple of hundred lines. With Motif you could cut it down to less than 50, but there's still significant set up. I don't have extensive VC++ experience but never recall it being quite that bad (of course, the wizards and GUI builders automated a lot of the process, and surely that would be considered a benefit?) -- Terry ###### From: Mike Swaim Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <01be2798$4885e7c0$159876c7@ga0235bbishop> <3676F14B.4E02ABDF@the_bottom_for_my.e-mail> <2230.653T2739T10516280@sky.bus.com> <75bvhq$evg$1@nntp1.u.washington.edu> Organization: PointeCom User-Agent: tin/pre-1.4-980818 ("Laura") (UNIX) (FreeBSD/2.2.8-RELEASE (i386)) Lines: 37 Message-ID: Date: Fri, 18 Dec 1998 17:03:06 GMT NNTP-Posting-Host: 209.127.0.130 X-Trace: news1.giganews.com 914000586 209.127.0.130 (Fri, 18 Dec 1998 11:03:06 CDT) NNTP-Posting-Date: Fri, 18 Dec 1998 11:03:06 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newspump.monmouth.com!newspeer.monmouth.com!nntp.giganews.com!news1.giganews.com.POSTED!gemini.c-com.net!swaim D. Peschel wrote: : In article , : Mike Swaim wrote: :>Charlie Gibbs wrote: :>: A fascinating alternate view. Thanks for pointing that out. :>: You've illustrated two totally different forms of abstraction. :>: And I'd say that's one of the most convincing arguments yet :>: that a system should provide both textual and graphic interfaces. : I don't know if Charlie meant, "provide them on two separate monitors at : the same time"! But I'm sure it's a useful feature, especially for : debugging, though I've never used it. :> AutoCad does that. The last place I worked at had a number of dual :>headed machines (one graphic monitor, one text monitor) for the CAD :>people. I suppose that with multimonitor support in Windows now, they :>could do it again. : Except they wouldn't do it as well as before. Windows' text support is : pretty uneven; I don't think it supports the standard features of VGA or : even EGA cards (viz., 64 colors of text, user-defined character sets, : 512-byte character sets, underlining). I believe that the current version of AutoCad is Windows only. It provides a command line in a text window. It should be simple enough to display that full screen on one monitor, with the drawing area on another one. AutoCad also supports putting commands on "buttons" (hot spots, really) on the pad or mouse (the mice attached to the pads typically had 8+ buttons) so the operator wouldn't have to take his hand off the graphics pad and use the keyboard. -- Mike Swaim, Avatar of Chaos: Disclaimer:I sometimes lie. Home: swaim@c-com.net Alum: swaim@alumni.rice.edu Quote: "Boingie"^4 Y,W&D ###### rom: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 18:03:02 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 25 Message-ID: <3679e74b.5891113@news.vip.net> References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.39 X-Trace: 914003625 A01OARAUVD427CCD1C usenet54.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.corridex.com!Supernews73!supernews.com!Supernews69!not-for-mail seebs@plethora.net (Peter Seebach) wrote: >In article <75aj3n$9hu$1@news.wlg.netlink.net.nz>, >Don Stokes wrote: >>I haven't had anything to do with the Sparc at assembly level, but I'm >>not all that surprised. > >This wasn't a Sparc; it was a 3/60. > >The content was a loop of copying 1,000,000 bytes of data from one array >to another and then back. It wasn't fast enough that I would think it had >optimized everything away. How could it? Unless it is known that the second array already was a copy of the first (You didn't specify this to be so.), the first copy has to be done. The second one can be optimized away MAYBE. Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <75aj3n$9hu$1@news.wlg.netlink.net.nz> <3679e74b.5891113@news.vip.net> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 19 Message-ID: Date: Fri, 18 Dec 1998 18:45:26 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 914006726 205.166.146.8 (Fri, 18 Dec 1998 12:45:26 CDT) NNTP-Posting-Date: Fri, 18 Dec 1998 12:45:26 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article <3679e74b.5891113@news.vip.net>, Gene Wirchenko wrote: >>The content was a loop of copying 1,000,000 bytes of data from one array >>to another and then back. It wasn't fast enough that I would think it had >>optimized everything away. > How could it? Unless it is known that the second array already >was a copy of the first (You didn't specify this to be so.), the first >copy has to be done. The second one can be optimized away MAYBE. Well, simple! It can notice that, after the loop, I ignore the contents of both arrays. :) -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 18 Dec 1998 22:34:36 +0100 Organization: My own Private Self Lines: 46 Sender: neil@chonsp.franklin.ch Message-ID: References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> X-Newsreader: Gnus v5.3/Emacs 19.34 don@news.daedalus.co.nz (Don Stokes) writes: > > C itself is perfectly capable of dealling with lengths and pointers. So long you do not mind having to supply 2 parameters for every string. > Where it falls down is the abortion that is the standard library, Yes. Because they tried to save having 2 parameters for each string. > which > in turn influence and are influenced by by Unix parameter passing > conventions, No. $ man 2 read NAME read - read from a file descriptor SYNOPSIS #include ssize_t read(int fd, void *buf, size_t count); This Unix kernel call clearly uses pointer (buf) and length (count), no zero termination in sight. So do all other system calls using strings that I know of. The zero terminated stuff is entirely the blame of the standard C library, and has infested nearly all other C libraries. > lack of ability to deal with length/pointer pairs as > single objects. Basically, C really just gives you the machine objects > and lets you have at it; Yes. And the kernal bites the bullet, the standard C library chickend out. -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 20:12:50 GMT Organization: . Lines: 74 Message-ID: <367a9dfd.830093@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> <3679f5fb.9052847@news.ricochet.net> <75duei$70a$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: pool02b-194-7-80-215.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-feed.inet.tele.dk!bofh.vszbr.cz!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach tsmurphy@cs.uiuc.edu (TERENCE MURPHY) on 18 Dec 1998 16:04:34 GMT to alt.folklore.computers: > In article <3679f5fb.9052847@news.ricochet.net>, > phil wrote: > > >Except that i thought that the argument was that NT was difficult to > >develop for, and i believe that it is not NT that is messy to program, > >but rather the Windows GUI. It's not so bad these days, but i'm still > >surprised about the amount of infrastructure that has to be put into > >place for even the simplest Windows app. > > Is there any GUI environment where you -don't- have to do this? A > decent "hello world" program in Xlib is a couple of hundred lines. > With Motif you could cut it down to less than 50, but there's still > significant set up. I don't have extensive VC++ experience but never Without using any wizard or GUI builder, it's basically something like: //-------------- #include int APIENTRY WinMain(HINSTANCE hInstance,HINSTANCE hPrevInstance, LPSTR lpCmdLine, int nCmdShow) { MessageBox("Hello, world", MB_OKONLY, "Hello", NULL); return 0; } //--------------- I don't have VC on this computer, so I may be mixing up the args a bit, but it's the general idea. A 'decent' hello world program OTOH: the sample that comes with VC creates its own window, has its own message pump, put shortly it does everything by hand - including things "Hello world" doesn't need to, like create a menu (complete with keyboard shortcuts), display an "about" box in a separate window, load strings from a resource module (requiring a declaration for storage and a load call per string) instead of just using quoted strings, etc. It's 191 lines long, of which I could trim at least 50 and still keep it readable. For example, all comments (some of which are only 2 words) are on lines of their own, where I usually fit the shorter ones in the right margin on code lines. Generally speaking I agree with you though. In windows (and using only MS products) you have three main choices: 1- do it the hard way 2- still use C, but use the MFC classes 3- use VB Nr 1 is too much work, and nrs 2 and 3 feature about the same compiled code size (including runtimes) and execution speed, where 3 is the easiest to develop by *very* far, but gives you less flexibility for the actual work behind the screens. That's why I tend to use a combination of 1 and 3: VB user interface, and the heavy work done by a DLL or an OCX in VC. This way my GUI apps are usually finished faster than they used to be in a CLI (read DOS), yet they have more functionality (meaning more & better ways to present data, more ways to do the same thing). -- News headlines. Earlier today, Saddam appealed onto all his countrymen to start a Jihad against men who cheat on their wives. ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 20:12:51 GMT Organization: . Lines: 14 Message-ID: <367ba8ab.3564455@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <36874dd4.3050706@news.uunet.be> <75dkgn$55a$2@ligarius.ultra.net> NNTP-Posting-Host: pool02b-194-7-80-215.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach jmfbahciv@aol.com on Fri, 18 Dec 98 12:50:33 GMT to alt.folklore.computers: > But the content has been very good for newbies (and even On that part I couldn't agree more (and those war fighters weren't as fanatical as I've seen before, so that part is softened a bit too.) -- News headlines. Earlier today, Saddam appealed onto all his countrymen to start a Jihad against men who cheat on their wives. ###### From: genew@vip.net (Gene Wirchenko) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 22:55:37 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 28 Message-ID: <367ab047.1047481@news.vip.net> References: <75aj3n$9hu$1@news.wlg.netlink.net.nz> <3679e74b.5891113@news.vip.net> Reply-To: genew@vip.net NNTP-Posting-Host: 204.209.212.46 X-Trace: 914021578 A01OARAUVD42ECCD1C usenet52.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.1/32.230 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!howland.erols.net!Supernews73!supernews.com!Supernews69!not-for-mail seebs@plethora.net (Peter Seebach) wrote: >In article <3679e74b.5891113@news.vip.net>, >Gene Wirchenko wrote: >>>The content was a loop of copying 1,000,000 bytes of data from one array >>>to another and then back. It wasn't fast enough that I would think it had >>>optimized everything away. > >> How could it? Unless it is known that the second array already >>was a copy of the first (You didn't specify this to be so.), the first >>copy has to be done. The second one can be optimized away MAYBE. > >Well, simple! It can notice that, after the loop, I ignore the contents >of both arrays. :) I suppose that is possible. That's the sort of thing you would put in a benchmark. This means we could have code that is very aggressively optimized for benchmarks. But don't we have that now? Sincerely, Gene Wirchenko Computerese Irregular Verb Conjugation: I have preferences. You have biases. He/She has prejudices. ###### Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? References: <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> Organization: Plethora . Net - more net, less spam! X-Newsreader: trn 4.0-test69 (20 September 1998) From: seebs@plethora.net (Peter Seebach) Lines: 26 Message-ID: Date: Fri, 18 Dec 1998 23:24:10 GMT NNTP-Posting-Host: 205.166.146.8 X-Trace: ptah.visi.com 914023450 205.166.146.8 (Fri, 18 Dec 1998 17:24:10 CDT) NNTP-Posting-Date: Fri, 18 Dec 1998 17:24:10 CDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!chippy.visi.com!news-out.visi.com!ptah.visi.com!not-for-mail In article , Neil Franklin wrote: >don@news.daedalus.co.nz (Don Stokes) writes: >> C itself is perfectly capable of dealling with lengths and pointers. >So long you do not mind having to supply 2 parameters for every string. Or doing them as objects - more overhead, but easier to read. >The zero terminated stuff is entirely the blame of the standard C >library, and has infested nearly all other C libraries. Yes. It's actually very convenient for many purposes, because you *almost* never need... >Yes. And the kernal bites the bullet, the standard C library chickend out. That's why I wrote a library for objects which look a lot like strings but can contain everything and know about their own lengths. -s -- Copyright 1998, All rights reserved. Peter Seebach / seebs@plethora.net C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon! Send me money - get cool programs and hardware! No commuting, please. Visit my new ISP --- More Net, Less Spam! ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Fri, 18 Dec 1998 23:36:54 -0600 Organization: MindSpring Enterprises Lines: 34 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <1dk6q0n.1oe4e2ldrk1uaN@n37-5.berlin.snafu.de> NNTP-Posting-Host: d1.56.d1.1a X-Server-Date: 19 Dec 1998 05:33:34 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newspump.monmouth.com!newspeer.monmouth.com!cpk-news-hub1.bbnplanet.com!news.gtei.net!firehose.mindspring.com!hshubs In article <1dk6q0n.1oe4e2ldrk1uaN@n37-5.berlin.snafu.de>, jnickelsen@acm.org (Juergen Nickelsen) wrote: >I wish I had known this when I ported a "make" program to VMS. Being >quite unexperienced then, I did not know where to look and figured out >the common success codes of the compiler and linker by trying. At least >that didn't take too long. Yeah. It looks like this: bits 31..16 15..0 facility code message code ...and bits 2..0 are the severity code. So if you mask out the high 16 bits, you can do quite a bit with the low 16 bits. If you mask out all but the bottom three bits, you'll do fine. It's just that you can get more specific when you use more bits. Using just bit zero will always tell you the most important ...er... bit. Most of the time, comparing with SS$_SUCCESS will work. The rest of the time, simply testing the status value as a logical/boolean will work as VMS treats an odd value as TRUE. So you can say (in FORTRAN): INTEGER*4 STATUS STATUS = SYS$FOO (...) IF (.NOT.STATUS) THEN END IF you get the idea. -- Howard S Shubs The Denim Adept ###### From: fleetvis@ricochet.net (phil) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 1998 09:07:12 GMT Organization: Posted via RemarQ, http://www.remarQ.com - Discussions start here! Lines: 32 Message-ID: <367b6647.14298780@news.ricochet.net> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n65-158.berlin.snafu.de> <367a265f.0@192.168.0.20> NNTP-Posting-Host: 204.179.130.33 X-Trace: 914058088 F7W8ZAFJM8221CCB3C usenet53.supernews.com X-Complaints-To: newsabuse@remarQ.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newsfeed.berkeley.edu!Supernews73!supernews.com!Supernews69!not-for-mail On Fri, 18 Dec 1998 09:50:34 -0000, "Samael" wrote: > >>phil wrote: >> >>> A GUI is a bitch to develop for. Of all those available, NT is probably >>> the easiest (you can produce a console app with minimal grief). > >Why is a GUI a bitch to develop for? > >I have programmed for GUI's and straight text app's and I find them equally >easy, just different. I think it depends on what you are trying to do. Just the fact that you can't output and forget can be a hurdle in many applications. If you use one of the existing controls it is hardly a printf() or a cout and it will eventually come and bite you when you least expect it. I think that the fact that many computer courses that are trying to teach C++ use either a Unix environment or a console application base says a great deal. I know many extremely good programmers who are intimidated by the GUI style of development and imho, quite rightly so. I have no life, so i gave up and wrote my own fire and forget classes that were derived from MFC stuff but provided the control i really needed including having to define a memory based compatible device context and a bitblt to avoid flickering, but i suspect that this was hardly germane to the "real" requirements. phil. ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec19.093804.15010@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> Lines: 49 Date: Sat, 19 Dec 1998 09:38:04 GMT NNTP-Posting-Host: 130.244.60.113 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 914060917 130.244.60.113 (Sat, 19 Dec 1998 10:48:37 MET DST) NNTP-Posting-Date: Sat, 19 Dec 1998 10:48:37 MET DST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article , Neil Franklin wrote: > don@news.daedalus.co.nz (Don Stokes) writes: > > C itself is perfectly capable of dealling with lengths and pointers. > So long you do not mind having to supply 2 parameters for every string. If you don't like two parameters then use a descriptor. typedef struct { size_t length; char *string; } string_descriptor; > $ man 2 read > > NAME > read - read from a file descriptor > SYNOPSIS > #include > ssize_t read(int fd, void *buf, size_t count); > > This Unix kernel call clearly uses pointer (buf) and length (count), > no zero termination in sight. > > So do all other system calls using strings that I know of. If you know about read(2) then you have to know about open(2) as well. | NAME | open - open for reading or writing | | C SYNOPSIS | #include | #include | #include | | int open (const char *path, int oflag, ... /* mode_t mode */); We also have link(2), unlink(2), stat(2), lstat(2), statvfs(2), chmod(2), chown(2), creat(2), mknod(2), utime(2) and many more. These system calls are part of the Unix heart so the Unix kernel is "infested" with zero terminated strings. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 98 13:01:16 GMT Organization: UltraNet Communications, Inc. Lines: 18 Message-ID: <75g9h2$dk3$1@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <36874dd4.3050706@news.uunet.be> <75dkgn$55a$2@ligarius.ultra.net> <367ba8ab.3564455@news.uunet.be> NNTP-Posting-Host: d12.dial-17.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 19 Dec 1998 13:25:54 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!feed1.news.rcn.net!rcn!news.ultranet.com!d12 In article <367ba8ab.3564455@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach jmfbahciv@aol.com on Fri, 18 Dec 98 12:50:33 GMT to >alt.folklore.computers: > >> But the content has been very good for newbies (and even > >On that part I couldn't agree more >(and those war fighters weren't as fanatical as I've seen before, > so that part is softened a bit too.) > Softened? Oh, no [emoticon hitting forehead]. Our design sessions were never soft :-). /BAH Subtract a hundred and four for e-mail. ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 1998 15:17:04 +0100 Organization: [Posted via] Interactive Networx Lines: 12 Message-ID: <1dk9cvg.5onksm162qy5kN@n246-123.berlin.snafu.de> References: <73ela0$51h@hermes.acs.unt.edu> <36874dd4.3050706@news.uunet.be> <75dkgn$55a$2@ligarius.ultra.net> <367ba8ab.3564455@news.uunet.be> NNTP-Posting-Host: n246-123.berlin.snafu.de User-Agent: MacSOUP/2.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.nacamar.de!news-hh.maz.net!unlisys!news.snafu.de!jnickelsen Luc Van der Veken wrote: > (and those war fighters weren't as fanatical as I've seen before, > so that part is softened a bit too.) Very much agreed. After getting interested because VMS was discussed, I was (and still am) very pleased how civilized and technical the discussion eveloved. I find the whole thread very interesting and instructive. -- Juergen Nickelsen ###### From: ejb@klamath.demon.co.uk (Edward John M. Brocklesby) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 1998 17:59:50 GMT Message-ID: <914090390.5872.0.nnrp-04.d4e4c6f2@news.demon.co.uk> References: <73ela0$51h@hermes.acs.unt.edu> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk> NNTP-Posting-Host: klamath.demon.co.uk X-NNTP-Posting-Host: klamath.demon.co.uk:212.228.198.242 X-Trace: news.demon.co.uk 914090390 nnrp-04:5872 NO-IDENT klamath.demon.co.uk:212.228.198.242 X-Complaints-To: abuse@demon.net X-Newsreader: slrn (0.9.4.3 UNIX) Lines: 31 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!bullseye.news.demon.net!demon!news.demon.co.uk!demon!klamath.demon.co.uk!ejb On about Wed, 09 Dec 1998 22:51:13 -0600, Howard S Shubs might have written: >In article <913227054.18505.0.nnrp-08.d4e4c6f2@news.demon.co.uk>, >ejb@klamath.demon.co.uk (Edward John M. Brocklesby) wrote: > >>We're porting our IRC bot (Ilona) to VMS at some point. (It might well be >>me who'll have to do this). But I can't find any documentation for it.. >>The system I'll be using has MULTINET TCP/IP installed, but I can't find >>anything like section 3 on Unix .. I suppose I just haven't looked hard >>enough :) Are there any manuals available online? > >No, neither are they available on UNIX. Far as I can tell, the only paper >manuals available for UNIX are compilations of MAN pages. There are no >actual manuals. I have next to be, a large set of binders labeled 'AT&T Software Library'. One binder, about 2", is the manpages (they aren't online). The rest is real life, useful manuals, describing how the system works. This was what I got with AT&T SVR2 .. I suspect it's very different with modern OS's. Such is the price of progress. >And note that VMS is being supplanted by Windows NT too. But we don't want to port it to NT :) >-- >Howard S Shubs The Denim Adept -- You will be misunderstood by everyone. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 1998 18:06:06 +0000 Organization: Tanglewood Message-ID: <367BEB0E.EA14B1CC@tnglwood.demon.co.uk> References: <75aj3n$9hu$1@news.wlg.netlink.net.nz> <3679e74b.5891113@news.vip.net> <367ab047.1047481@news.vip.net> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 914092683 nnrp-04:7259 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 18 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!feed1.news.rcn.net!rcn!woodstock.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Gene Wirchenko wrote: > put in a benchmark. This means we could have code that is very > aggressively optimized for benchmarks. But don't we have that now? > I seem to remember that when 386 PCs were new, someone who should have known better quoted the Whetstone figures to argue that they were faster than VAXen. I constructed a benchmark which had to print a number (the sum of the first 1000 primes IIRC) to prove that it had really done the calculation, not just optimised it away. Suddenly the 386 was *very* slow beside the VAX. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 19 Dec 1998 20:54:08 +0100 Organization: My own Private Self Lines: 28 Sender: neil@chonsp.franklin.ch Message-ID: References: <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> <1998Dec19.093804.15010@lorelei.approve.se> X-Newsreader: Gnus v5.3/Emacs 19.34 hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) writes: > > Neil Franklin wrote: > > ssize_t read(int fd, void *buf, size_t count); > > > > This Unix kernel call clearly uses pointer (buf) and length (count), > > no zero termination in sight. > > > > So do all other system calls using strings that I know of. > > If you know about read(2) then you have to know about open(2) as well. > > | int open (const char *path, int oflag, ... /* mode_t mode */); Interesting. I must have know that 10 years ago when I did a lot of C application programming... I suppose that always using read(1, ...) and write(2, ...) (or fread(stdin, ...) and fwrite(stdout, ...) since getting onto Unix sort of lets one forget stuff like that :-). -- Neil Franklin, Nerd, Geek, Unix Guru, Hacker, Mystic neil@franklin.ch.remove http://neil.franklin.ch/ "No, it was a JOKE! You can't RUN this!" Ken Thompson ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sat, 19 Dec 1998 19:50:05 GMT Organization: . Lines: 76 Message-ID: <367f023e.5991525@news.uunet.be> References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: pool02b-194-7-84-162.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!newsfeed.wirehub.nl!btnet-peer!btnet!rill.news.pipex.net!pipex!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach Neil Franklin on 18 Dec 1998 22:34:36 +0100 to alt.folklore.computers: > don@news.daedalus.co.nz (Don Stokes) writes: > > > > C itself is perfectly capable of dealling with lengths and pointers. > > So long you do not mind having to supply 2 parameters for every string. Isn't this what structs were designed for? (see examples below) > NAME > read - read from a file descriptor > SYNOPSIS > #include > ssize_t read(int fd, void *buf, size_t count); > > This Unix kernel call clearly uses pointer (buf) and length (count), > no zero termination in sight. When did 'read' become a string handling function????? I use it for reading binary data - including zeroes - from files. > > So do all other system calls using strings that I know of. So the unix kernel people didn't use structs (at least for strings). Does that mean we can't use them in our own code? [As you can tell by my comments,] I've been using a C++ compiler to compile my C code for so long that I don't remember if C allows a typedef directly on a struct this way, but: [I never tried to compile even a single line of this, so please nobody come complaining if...] typedef struct { size_t length; char *text; // Space allocated at runtime } MY_STRING, *pMY_STRING; int print_my_string(pMY_STRING); MY_STRING x; x.length = strlen(sourcestring); x.text = malloc(x.length); memmove(x.text, sourcestring, x.length); ... print_my_string(&x); In C++ you'd make it a class and add access members to make it easier to handle. Using a union allows you to drop one level of indirection and work with 'real' size-prefixed strings: typedef union { // Total space allocated at runtime size_t length; char dontuse[sizeof(size_t)], data[]; } *pMyString; pMyString pstr; pstr = malloc(strlen(sourcestring) + sizeof(size_t)); pstr->length = strlen(sourcestring); memmove(pstr->data, sourcestring, pstr->length); BTW - not that I think anyone in this group is stupid, but - if you're wondering why I used memmove instead of strcpy, think of efficiency: 4 bytes at a time on a 32 bit machine, repxx instructions, plus no longer the overhead of having to check each character for being the terminating zero. (yes, I know I called strlen twice on the same string :-) ###### From: don@news.daedalus.co.nz (Don Stokes) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 19 Dec 1998 19:59:04 GMT Organization: Daedalus Consulting Lines: 18 Message-ID: <75h0i8$3mv$1@news.wlg.netlink.net.nz> References: <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> NNTP-Posting-Host: toyunix.zl2tnm.gen.nz Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!203.97.37.7!newsfeed.clear.net.nz!news.wlg.netlink.net.nz!don Neil Franklin wrote: >don@news.daedalus.co.nz (Don Stokes) writes: >> C itself is perfectly capable of dealling with lengths and pointers. >So long you do not mind having to supply 2 parameters for every string. Or passing a pointer to a descriptor, like the VMS libraries do. >> in turn influence and are influenced by by Unix parameter passing >> conventions, >This Unix kernel call clearly uses pointer (buf) and length (count), >no zero termination in sight. Gee, I *distinctly* recall null-terminating filenames for open(), rename() & friends. Must have been hallucinating. Funny, I can't find a length parameter in the man pages though. -- Don Stokes, Networking Consultant http://www.daedalus.co.nz +64 25 739 724 ###### Sender: Torsten Poulin From: Torsten Poulin Subject: Re: Is VMS taboo, or what? Newsgroups: alt.folklore.computers References: <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> Date: Sun, 20 Dec 1998 01:47:44 +0100 Message-ID: NNTP-Posting-Host: 194.19.185.98 X-NNTP-Posting-Host: 194.19.185.98 X-Trace: 20 Dec 1998 01:47:43 -0100, 194.19.185.98 Lines: 28 X-Complaints-To: abuse@telia.dk Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!newsfeed101.telia.com!d2o105.telia.com!not-for-mail Peter Seebach skrev: > This wasn't a Sparc; it was a 3/60. > > The content was a loop of copying 1,000,000 bytes of data from one array > to another and then back. It wasn't fast enough that I would think it had > optimized everything away. Ahh, a 68k? Then it could have been something along the lines of using ... MOVEM.L source,D0-D7/A0-A7 MOVEM.L D0-D7/A0-A7,destination ADDI #64,source ADDI #64,destination ... to move 64 bytes at a time from the source into the registers and then onto the destination. Definitely a lot faster than moving individual bytes in a loop, even when we have to allow for special casing the ends which may not be on longword boundaries and saving and restoring the context on entry and exit. I may have the details wrong (it's been a long time), but you get the idea. -Torsten ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 20 Dec 1998 02:41:03 GMT Lines: 26 Message-ID: <75ho3v$vrm$2@roch.zetnet.co.uk> References: <367f023e.5991525@news.uunet.be> NNTP-Posting-Host: man-087.dialup.zetnet.co.uk X-Trace: roch.zetnet.co.uk 914121663 32630 194.247.41.108 (20 Dec 1998 02:41:03 GMT) NNTP-Posting-Date: 20 Dec 1998 02:41:03 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!195.200.0.51.MISMATCH!newshub.bart.net!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-19 lucvdv@null.net(LucVanderVeken) said: :> So long you do not mind having to supply 2 parameters for every :string. :Isn't this what structs were designed for? (see examples below) K&R C didn't allow returning a struct from a function (or passing it in by value either). :BTW - not that I think anyone in this group is stupid, but - if :you're wondering why I used memmove instead of strcpy, think of :efficiency: 4 bytes at a time on a 32 bit machine, repxx :instructions, plus no longer the overhead of having to check each :character for being the terminating zero. :(yes, I know I called strlen twice on the same string :-) Which will almost certainly wipe out any efficiency gains: 1. code / time out the results of 'rep movsd' on a 486 vs the alternative "obvious" code - you'll be surprised, I was; 2. bear in mind that strlen() will take the same order of time as a strcpy() will - you'll make a net loss on this code. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 03:22:46 +0100 Organization: [Posted via] Interactive Networx Lines: 23 Message-ID: <1dkak5n.1owzq0hp5fggaN@n164-118.berlin.snafu.de> References: <366542cc.0@news.mtx.net.au> <912721959.24008.0.nnrp-07.d4e4c6f2@news.demon.co.uk> <749lip$gjf$1@vixen.cso.uiuc.edu> <912986787.12615.0.nnrp-01.9e9878e0@news.demon.co.uk> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74h82f$4mt@hermes.acs.unt.edu> <1998Dec8.101907.3786@lorelei.approve.se> <821.647T270T6923295@sky.bus.com> <755ubm$quj$1@irk.zetnet.co.uk> <2uFd2.2296$WZ6.6110370@ptah.visi.com> <368118f0.12180594@news.uunet.be> <75aj3n$9hu$1@news.wlg.netlink.net.nz> <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> <1998Dec19.093804.15010@lorelei.approve.se> NNTP-Posting-Host: n164-118.berlin.snafu.de User-Agent: MacSOUP/2.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.tli.de!unlisys!news.snafu.de!jnickelsen Goran Larsson wrote: [Strings] > If you don't like two parameters then use a descriptor. > > typedef struct { > size_t length; > char *string; > } string_descriptor; I'd consider having typedef struct { size_t length ; /* Actually used part of string */ size_t allocated_length ; /* Size of allocated space at string */ char *string ; } string_descriptor ; Preallocating fixed sizes can save reallocations for very dynamic strings and speed up alloc/free operations. -- Juergen Nickelsen ###### From: hshubs@mindspring.com (Howard S Shubs) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 08:52:12 -0600 Organization: MindSpring Enterprises Lines: 8 Message-ID: References: <73ela0$51h@hermes.acs.unt.edu> <365F8990.12A351F8@nospam.usask.ca> <367dd2f8.2096484@news.uunet.be> NNTP-Posting-Host: d1.56.d2.fe X-Server-Date: 20 Dec 1998 14:52:13 GMT X-Newsreader: MT-NewsWatcher 2.4.4 X-Face: "S"r{U%bs].&Ud}Pc~~~0a]M:t5l>>EN\1Faw10M9NK1Xq59wo7-"s0S+[{etQorO /Nf-Ci"i9v'MT!R8)J]N[4|2&x1r^Iq&{SB"6dknr0=+6UFb.>+{zMn_1=rw&/V+"d@* ZS5\LoW_ Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!firehose.mindspring.net!firehose.mindspring.com!hshubs In article <367dd2f8.2096484@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >With Compaq at the wheel, maybe VMS 8.0 will be written by MS? AEEEEIIIIIIIIEEIEEIEIEIEIIEIEIEEEEE!!!!!!!! -- Howard S Shubs The Denim Adept ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 11:38:13 GMT Organization: . Lines: 27 Message-ID: <367ccd1e.598560@news.uunet.be> References: <367f023e.5991525@news.uunet.be> <75ho3v$vrm$2@roch.zetnet.co.uk> NNTP-Posting-Host: pool02b-194-7-41-164.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach lisard@zetnet.co.uk on 20 Dec 1998 02:41:03 GMT to alt.folklore.computers: > :(yes, I know I called strlen twice on the same string :-) > Which will almost certainly wipe out any efficiency gains: That smiley meant "I should have stored the result in a temporary variable, but I was too lazy when putting this example toghether" > 1. code / time out the results of 'rep movsd' on a 486 vs the > alternative "obvious" code - you'll be surprised, I was; The CPU doesn't have to read a new instruction in between every two dword moves, so I'd expect it to be at least 50% faster. Some posts back in this thread the discussion was about the huge difference in speed between memmove and a 'for' loop in plain C. I couldn't believe it made a 280 to 1 difference, but it was claimed to be that high (on an ALPHA, not on Intel silicon). > 2. bear in mind that strlen() will take the same order of time > as a strcpy() will - you'll make a net loss on this code. I don't have a C compiler at hand to check the libraries, but I think the actual search for 0 can be done in a single assembler instruction (rep scasb?), which would be a lot faster. ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 11:38:15 GMT Organization: . Lines: 42 Message-ID: <367dd2f8.2096484@news.uunet.be> References: <73ela0$51h@hermes.acs.unt.edu> <365F8990.12A351F8@nospam.usask.ca> NNTP-Posting-Host: pool02b-194-7-41-164.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!diablo.theplanet.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach Chris Pitzel on Fri, 27 Nov 1998 23:26:40 -0600 to alt.folklore.computers: > If DEC had marketed VMS properly, we'd all be sitting infront of VMS > running machines today (or at least a good number of us.). Over time, > they would've reworked the GUI into something nice and pretty (like > WindowsNT is today..), added a working and half-decent POSIX subsystem, > and added in all of the multimedia components and networking prowess and > ease of use that is expected of an OS today. Aren't we in a way sitting infront of VMS with MS's GUI running as the shell? Yesterday, in a magazine stand, I noticed Windows NT magazine's december issue features a cover story "Is NT Windows?", with articles "NT and VMS" and "NT vs. Unix" (both by Mark Russinovich). The VMS article explains NT's VMS roots, and also that MS payed Digital somewhere between 65 and 100 million $$ and helped Digital train NT technicians, in return for not getting sued. In civilized language, this sounds like "Digital cut a deal with MS." :-) These lines near the end sound interesting: " ... throughout the 1990s, Digital introduced many NT features to VMS, and Microsoft has added VMS development to NT. For example, VMS featured native clustering support in 1984, and 64-bit memory and system APIs in 1996. Microsoft did not introduce clustering support to NT until late last year - and only on a limited scale - ... VMS introduced kernel-mode threads in VMS 7.0 in 1995, and VMS 7.2 will include NT-style event logging and a Registry." With Compaq at the wheel, maybe VMS 8.0 will be written by MS? -- [If they sue me for quoting here, next time I'll change the wording a bit and forget to advertize for them] ###### Newsgroups: alt.folklore.computers From: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Subject: Re: Is VMS taboo, or what? Message-ID: <1998Dec20.122132.24769@lorelei.approve.se> Originator: hoh@approve.se.NO_JUNK_EMAIL (Goran Larsson) Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & X-No-Archive: yes X-Newsreader: trn 4.0-test66 (4 June 1998) References: <367f023e.5991525@news.uunet.be> <75ho3v$vrm$2@roch.zetnet.co.uk> Lines: 14 Date: Sun, 20 Dec 1998 12:21:32 GMT NNTP-Posting-Host: 130.244.75.72 X-Complaints-To: news-abuse@swip.net X-Trace: nntpserver.swip.net 914156919 130.244.75.72 (Sun, 20 Dec 1998 13:28:39 MET DST) NNTP-Posting-Date: Sun, 20 Dec 1998 13:28:39 MET DST Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!nntpserver.swip.net!not-for-mail In article <75ho3v$vrm$2@roch.zetnet.co.uk>, wrote: > K&R C didn't allow returning a struct from a function (or passing it in > by value either). That is true for the C described in the first edition of the book "The C Programming Language" but this part of C was added just after that book was written. I don't know of any C compiler that didn't allow this. -- Goran Larsson hoh AT approve DOT se I was an atheist, http://home DOT swipnet DOT se/hoh/ until I found out I was God. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 98 13:12:28 GMT Organization: UltraNet Communications, Inc. Lines: 45 Message-ID: <75iuid$i49$2@strato.ultra.net> References: <73ela0$51h@hermes.acs.unt.edu> <365F8990.12A351F8@nospam.usask.ca> <367dd2f8.2096484@news.uunet.be> NNTP-Posting-Host: d13.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 20 Dec 1998 13:37:17 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!newshub.northeast.verio.net!newsfeed.xcom.net!news.ultranet.com!d13 In article <367dd2f8.2096484@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach Chris Pitzel on Fri, >27 Nov 1998 23:26:40 -0600 to alt.folklore.computers: > >> If DEC had marketed VMS properly, we'd all be sitting infront of VMS >> running machines today (or at least a good number of us.). Over time, >> they would've reworked the GUI into something nice and pretty (like >> WindowsNT is today..), added a working and half-decent POSIX subsystem, >> and added in all of the multimedia components and networking prowess and >> ease of use that is expected of an OS today. > >Aren't we in a way sitting infront of VMS with MS's GUI running >as the shell? > >Yesterday, in a magazine stand, I noticed Windows NT magazine's >december issue features a cover story "Is NT Windows?", with >articles "NT and VMS" and "NT vs. Unix" (both by Mark >Russinovich). Who is he? > >The VMS article explains NT's VMS roots, and also that MS payed >Digital somewhere between 65 and 100 million $$ and helped >Digital train NT technicians, in return for not getting sued. I hadn't heard about the "getting sued" part. But JMF was one who got assiged the NT effort. He was just a tad more than a technician. I don't know what the project plan said..I just know that he got some of the work. >In civilized language, this sounds like "Digital cut a deal with >MS." :-) The deal cutting was highly publicized and deemed a "feature" for the business. I had my own hypothesis, but it was based on no facts. /BAH Subtract a hundred and four for e-mail. ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 98 13:19:52 GMT Organization: UltraNet Communications, Inc. Lines: 36 Message-ID: <75iv08$i49$3@strato.ultra.net> References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> <367f023e.5991525@news.uunet.be> NNTP-Posting-Host: d13.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 20 Dec 1998 13:44:40 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news.idt.net!newspeer.monmouth.com!news.ultranet.com!d13 In article <367f023e.5991525@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach Neil Franklin on 18 Dec >1998 22:34:36 +0100 to alt.folklore.computers: > >> don@news.daedalus.co.nz (Don Stokes) writes: >> > >> > C itself is perfectly capable of dealling with lengths and pointers. >> >> So long you do not mind having to supply 2 parameters for every string. > >Isn't this what structs were designed for? (see examples below) > >BTW - not that I think anyone in this group is stupid, but - if >you're wondering why I used memmove instead of strcpy, think of >efficiency: 4 bytes at a time on a 32 bit machine, repxx >instructions, plus no longer the overhead of having to check each >character for being the terminating zero. >(yes, I know I called strlen twice on the same string :-) > Well, I'm stupid (or uneducated). The drift of this thread is implying that data transfer by the processor is byte based? And this is called RISC? This doesn't make any sense. If an application asks the processor to move some stuff, why should the processor care what form it's in? And why does everything seem to be in strings? Bletcho! Just showing my old slip :-). /BAH Subtract a hundred and four for e-mail. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 14:22:08 +0000 Organization: Tanglewood Message-ID: <367D0810.60F8F401@tnglwood.demon.co.uk> References: <367f023e.5991525@news.uunet.be> <75ho3v$vrm$2@roch.zetnet.co.uk> <1998Dec20.122132.24769@lorelei.approve.se> NNTP-Posting-Host: localhost.demon.co.uk X-NNTP-Posting-Host: tnglwood.demon.co.uk:158.152.132.30 X-Trace: news.demon.co.uk 914164417 nnrp-01:2410 NO-IDENT tnglwood.demon.co.uk:158.152.132.30 X-Complaints-To: abuse@demon.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.07 [en] (X11; I; Linux 2.0.31 i586) Lines: 13 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news.demon.co.uk!demon!tnglwood.demon.co.uk!not-for-mail Goran Larsson wrote: > that book was written. I don't know of any C compiler that didn't > allow this. I don't thing Whitesmiths did, but hardly anything else worked in Whitesmiths anyway, so it didn't matter. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: lucvdv@null.net (Luc Van der Veken) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Sun, 20 Dec 1998 19:16:52 GMT Organization: . Lines: 30 Message-ID: <36824934.6360325@news.uunet.be> References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> <367f023e.5991525@news.uunet.be> <75iv08$i49$3@strato.ultra.net> NNTP-Posting-Host: pool02b-194-7-147-75.uunet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 X-No-Archive: yes Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!newsfeed.ecrc.net!join.news.pipex.net!pipex!krypton.inbe.net!INbe.net!not-for-mail Also sprach jmfbahciv@aol.com on Sun, 20 Dec 98 13:19:52 GMT to alt.folklore.computers: > The drift of this thread > is implying that data transfer by the processor is byte based? That's not how I would put it. The drift is toward how C works with character strings (we seem to be losing sight of the original subject). > And this is called RISC? Not entirely - I kinda dragged in Intel too (with the rep xxx), generalizing toward all 32-bit processors. > This doesn't make any sense. If > an application asks the processor to move some stuff, why > should the processor care what form it's in? I hope it doesn't. But moving 4 bytes one byte at a time will be slower than moving the 4 of them in a single 32-bit transfer, on any processor. > And why > does everything seem to be in strings? Bletcho! > > Just showing my old slip :-). Whell, most strings I know of are in strings most of the time. But there's a significant difference between the C and the G kind :) ###### Newsgroups: alt.folklore.computers Date: Sun, 20 Dec 1998 23:49:12 +0100 Message-ID: From: peterk@combo.ganesha.com (Dr. Peter Kittel) Subject: Re: Is VMS taboo, or what? Reply-To: peterk @ combo.ganesha.com References: <73ela0$51h@hermes.acs.unt.edu> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> <3679f5fb.9052847@news.ricochet.net> <75duei$70a$1@vixen.cso.uiuc.edu> X-Newsreader: rn7.bas Lines: 23 Organization: Private Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!btnet-peer!btnet!news-lond.gip.net!news.gsl.net!gip.net!newsfeed.ecrc.net!blackbush.xlink.net!ganesha.ganesha.com!shorter!combo.ganesha.com!peterk In article <75duei$70a$1@vixen.cso.uiuc.edu> tsmurphy@cs.uiuc.edu (TERENCE MURPHY) writes: >In article <3679f5fb.9052847@news.ricochet.net>, >phil wrote: > >>Except that i thought that the argument was that NT was difficult to >>develop for, and i believe that it is not NT that is messy to program, >>but rather the Windows GUI. It's not so bad these days, but i'm still >>surprised about the amount of infrastructure that has to be put into >>place for even the simplest Windows app. > >Is there any GUI environment where you -don't- have to do this? Sure. > A >decent "hello world" program in Xlib is a couple of hundred lines. Omegod. Run and have a look at AmigaOS or more recent BeOS. A dozen lines, perhaps two dozen maximum. -- Best Regards, Dr. Peter Kittel // E-Mail: Private Site in Frankfurt, Germany \X/ peterk @ combo.ganesha.com ###### From: "The XO" Newsgroups: alt.folklore.computers References: <73ela0$51h@hermes.acs.unt.edu> <365F8990.12A351F8@nospam.usask.ca> <367dd2f8.2096484@news.uunet.be> Subject: Re: Is VMS taboo, or what? Date: Mon, 21 Dec 1998 11:09:58 +1030 Lines: 20 Organization: Headquarters X-Newsreader: Microsoft Outlook Express 4.71.1712.3 X-Mimeole: Produced By Microsoft MimeOLE V4.71.1712.3 NNTP-Posting-Host: netcafe.pirie.mtx.net.au Message-ID: <367d9889.0@news.mtx.net.au> X-Trace: 21 Dec 1998 11:08:33 -1050, netcafe.pirie.mtx.net.au Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!harbinger.cc.monash.edu.au!news.uwa.edu.au!news1.optus.net.au!optus!yorrell.saard.net!duster.adelaide.on.net!news.saix.saia.asn.au!news.mtx.net.au!netcafe.pirie.mtx.net.au Luc Van der Veken wrote in message <367dd2f8.2096484@news.uunet.be>... >With Compaq at the wheel, maybe VMS 8.0 will be written by MS? Shudder!!! Wash your mouth out with soap!!! Nightmares are made of that. Cheers Geoff Roberts Computer Room Internet Cafe Port Pirie South Australia Spam countermeasures netcafe at pirie dot mtx dot net dot au ###### From: "Samael" Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 21 Dec 1998 13:06:14 -0000 Lines: 30 Message-ID: <367e47d6.0@192.168.0.20> References: <73ela0$51h@hermes.acs.unt.edu> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> <3679f5fb.9052847@news.ricochet.net> <75duei$70a$1@vixen.cso.uiuc.edu> NNTP-Posting-Host: 195.147.134.72 X-Newsreader: Microsoft Outlook Express 4.72.3155.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet!nntp.news.xara.net!xara.net!news.itg.net.uk!192.168.0.20!192.168.0.1 TERENCE MURPHY wrote in message <75duei$70a$1@vixen.cso.uiuc.edu>... >In article <3679f5fb.9052847@news.ricochet.net>, >phil wrote: > >>Except that i thought that the argument was that NT was difficult to >>develop for, and i believe that it is not NT that is messy to program, >>but rather the Windows GUI. It's not so bad these days, but i'm still >>surprised about the amount of infrastructure that has to be put into >>place for even the simplest Windows app. > >Is there any GUI environment where you -don't- have to do this? A >decent "hello world" program in Xlib is a couple of hundred lines. >With Motif you could cut it down to less than 50, but there's still >significant set up. I don't have extensive VC++ experience but never >recall it being quite that bad (of course, the wizards and GUI builders >automated a lot of the process, and surely that would be considered a >benefit?) Lets see, create a 'form', add a button to it, and a text box. In the click method of the button, set the text box value to "Hello World". Code: 1 line. Time taken - about 1 minute (including the compile ,f course). Samael ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Mon, 21 Dec 98 14:19:40 GMT Organization: UltraNet Communications, Inc. Lines: 54 Message-ID: <75lmsq$gbk$1@strato.ultra.net> References: <75bb78$n29$1@vixen.cso.uiuc.edu> <75bmfo$59g$1@news.wlg.netlink.net.nz> <75btgp$b6i$1@vixen.cso.uiuc.edu> <75c5cl$cg7$1@news.wlg.netlink.net.nz> <367f023e.5991525@news.uunet.be> <75iv08$i49$3@strato.ultra.net> <36824934.6360325@news.uunet.be> NNTP-Posting-Host: d1.dial-14.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 21 Dec 1998 14:44:42 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!feed1.news.luth.se!luth.se!cam-news-hub1.bbnplanet.com!news.gtei.net!newsfeed.xcom.net!news.ultranet.com!d1 In article <36824934.6360325@news.uunet.be>, lucvdv@null.net (Luc Van der Veken) wrote: >Also sprach jmfbahciv@aol.com on Sun, 20 Dec 98 13:19:52 GMT to >alt.folklore.computers: > >> The drift of this thread >> is implying that data transfer by the processor is byte based? > >That's not how I would put it. The drift is toward how C works >with character strings (we seem to be losing sight of the >original subject). Well, that happens in all meetings...thread drift is just an emulation. > >> And this is called RISC? >Not entirely - I kinda dragged in Intel too (with the rep xxx), >generalizing toward all 32-bit processors. >> This doesn't make any sense. If >> an application asks the processor to move some stuff, why >> should the processor care what form it's in? > >I hope it doesn't. But moving 4 bytes one byte at a time will be >slower than moving the 4 of them in a single 32-bit transfer, on >any processor. Sometimes. It does depend what the mover wants to do with them. > >> And why >> does everything seem to be in strings? Bletcho! >> >> Just showing my old slip :-). > >Whell, most strings I know of are in strings most of the time. I suppose, if one wants to get picky, one could consider bit moving a string of bits. I seem to remember a heated dicussion about that very thing in the late 70s. I guess the key is determining the criteria that stops the data trasfer. >But there's a significant difference between the C and the G kind :) Chuckle. OOhhhh, you're bad. From what I've heard one kind only works well in hot spots. And the other kind has to be cooled. /BAH Subtract a hundred and four for e-mail. ###### From: tsmurphy@cs.uiuc.edu (TERENCE MURPHY) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 21 Dec 1998 17:12:59 GMT Organization: University of Illinois at Urbana-Champaign Lines: 13 Message-ID: <75lvir$b3t$1@vixen.cso.uiuc.edu> References: <367f023e.5991525@news.uunet.be> <75ho3v$vrm$2@roch.zetnet.co.uk> <367ccd1e.598560@news.uunet.be> NNTP-Posting-Host: sparc31.cs.uiuc.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news.tele2.nl!newsfeed1.swip.net!swipnet!newsfeed.cwix.com!128.174.5.49!vixen.cso.uiuc.edu!tsmurphy In article <367ccd1e.598560@news.uunet.be>, Luc Van der Veken wrote: >I don't have a C compiler at hand to check the libraries, but I >think the actual search for 0 can be done in a single assembler >instruction (rep scasb?), which would be a lot faster. "rep scasb" is not a single assembler instruction, well, it loops each time it does that. Using the rep statement is no faster than using jnz to go back to the top (I'm sure rep is decoded into this in microcode). -- Terry ###### From: benh@lsl.co.uk (Ben Hutchings) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 21 Dec 1998 17:13:46 GMT Organization: Laser-Scan Ltd. Lines: 12 Message-ID: <75lvka$am9@relay.lsl.co.uk> References: NNTP-Posting-Host: 192.9.200.20 X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!diablo.theplanet.net!join.news.pipex.net!pipex!news.lsl.co.uk!benh Neil Franklin (neil@franklin.ch.remove) wrote: : This Unix kernel call clearly uses pointer (buf) and length (count), : no zero termination in sight. : So do all other system calls using strings that I know of. Don't you know the function open()? ;-) -- Any opinions expressed are my own and not necessarily those of Laser-Scan. ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: 21 Dec 1998 19:13:42 GMT Lines: 74 Message-ID: <75m6l6$guo$4@roch.zetnet.co.uk> References: <367ccd1e.598560@news.uunet.be> NNTP-Posting-Host: man-140.dialup.zetnet.co.uk X-Trace: roch.zetnet.co.uk 914267622 17368 194.247.40.178 (21 Dec 1998 19:13:42 GMT) NNTP-Posting-Date: 21 Dec 1998 19:13:42 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-12-20 lucvdv@null.net(LucVanderVeken) said: :> 1. code / time out the results of 'rep movsd' on a 486 vs the :> alternative "obvious" code - you'll be surprised, I was; :The CPU doesn't have to read a new instruction in between every :two dword moves, so I'd expect it to be at least 50% faster. Just checked with the list of timings I have. MOVSD ; 7 cycles/repetition cycle: MOV EAX, [ESI] MOV [EDI], EAX ADD SI, 4 ADD DI, 4 LOOP cycle ; 10 cycles/rep (LOOP is 6 on 486) less 4 cycles cycle2: MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX DEC ECX JNZ cycle2 ; 6 cycles/rep (JNZ is 3) less 2 cycles cycle3: MOV EAX, ECX AND EAX, 7 JMP [EAX*4+jmps] jmps: ALIGN 4 DD OFFSET st0, OFFSET st0 + 42, OFFSET st0 + 36, OFFSET st0 + 30 DD OFFSET st0 + 24, OFFSET st0 + 18, OFFSET st0 + 12, OFFSET st0 + 6 st0: MOV EAX, [ESI+ECX*4] \ 6 bytes MOV [EDI+ECX*4], EAX / 2 cycles MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX MOV EAX, [ESI+ECX*4] MOV [EDI+ECX*4], EAX SUB ECX, 8 JNZ cycle3 ; 2.5 cycles/rep (JNZ is 3) plus one JMP [X] Which is a slashing of nearly 3 over the simple alternative (and it will work on all the x86s, nearly!). Unrolling the loop even further gets you even closer to the possible 2 cycles per transfer at top speed. (To get to this point on a Pentium you'll probably want to double up the register usage, loading into EAX & EBX & then storing from both.) :Some posts back in this thread the discussion was about the huge :difference in speed between memmove and a 'for' loop in plain C. :I couldn't believe it made a 280 to 1 difference, but it was :claimed to be that high (on an ALPHA, not on Intel silicon). Well, the Alpha is 64 bit for a start, so transferring a byte at a time will get you an 8-fold slowdown before you consider such things as register use, the fact of using array arithmetic rather than pointer arithmetic, etc.etc. -- Communa (lisard@zetnet.co.uk) -- you know soft spoken changes nothing Net-Tamer V 1.08X - Test Drive ###### From: David Wragg Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Lines: 68 Message-ID: References: <367ccd1e.598560@news.uunet.be> <75m6l6$guo$4@roch.zetnet.co.uk> Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII X-Newsreader: Gnus v5.5/XEmacs 20.4 - "Emerald" Date: 22 Dec 1998 00:59:42 +0000 NNTP-Posting-Host: 194.119.176.228 X-Complaints-To: news@u-net.net X-Trace: newsr2.u-net.net 914299830 194.119.176.228 (Tue, 22 Dec 1998 04:10:30 BST) NNTP-Posting-Date: Tue, 22 Dec 1998 04:10:30 BST Organization: (Posted via) U-NET Internet Ltd. Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!news-feed1.eu.concert.net!btnet-peer!btnet!peer.news.th.u-net.net!u-net!newsr2.u-net.net.POSTED!gatsby.u-net.com!not-for-mail lisard@zetnet.co.uk writes: > On 1998-12-20 lucvdv@null.net(LucVanderVeken) said: > :> 1. code / time out the results of 'rep movsd' on a 486 vs the > :> alternative "obvious" code - you'll be surprised, I was; > > :The CPU doesn't have to read a new instruction in between every > :two dword moves, so I'd expect it to be at least 50% faster. > > Just checked with the list of timings I have. > > [snip assembly] > > Which is a slashing of nearly 3 over the simple alternative (and it > will work on all the x86s, nearly!). Unrolling the loop even further > gets you even closer to the possible 2 cycles per transfer at top speed. Yes, on the 486 rep movsd is particularly slow compared to simple assembly. (Though not on the 386) > (To get to this point on a Pentium you'll probably want to double up the > register usage, loading into EAX & EBX & then storing from both.) But you can't really generalize between x86 generations. On the Pentium, the fastest way is to use the FPU to do 64-bit reads and writes (or MMX these days). In addition, taking the cache line size and cache behaviour is very important to get the highest performance when the reads and writes have to go to main memory. But on the P6 processors, rep movsd becomes an effective method again. Since the P6 core executes RISC-like uops, rather than x86 instructions themselves, so the instruction is microcoded to simply generate an efficient sequence of uops for the task. > :Some posts back in this thread the discussion was about the huge > :difference in speed between memmove and a 'for' loop in plain C. > :I couldn't believe it made a 280 to 1 difference, but it was > :claimed to be that high (on an ALPHA, not on Intel silicon). > > Well, the Alpha is 64 bit for a start, so transferring a byte at a time > will get you an 8-fold slowdown before you consider such things as > register use, the fact of using array arithmetic rather than pointer > arithmetic, etc.etc. The original Alpha (21064) doesn't have instruction to access memory a byte at a time (even though it is byte addressed), so a byte at a time copy would be insane: read in a source word, *read* in the destination word, do a few instructions of bit twiddling, write the destination word and repeat. Far more than an 8-fold slowdown over how it is actually coded. Of course how it should work is that if the compiler sees something like: int i; char *src, *dst; for (i = 0; i < N; i += STRIDE) src[i] = dst[i]; then it does something clever if STRIDE < 8. A dumb compiler probably explains the 1:280 figure someone produced. Later Alpha implementations extended the instruction set with memory accesses for less than a whole word. I have heard that the main motivation for this is to do with memory page protection: Imagine you have a page marked write-only, and you want to write just one byte... Dave Wragg ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 05 Jan 1999 17:50:26 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 41 Message-ID: <36924ec2.185632234@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <3662237d.19279614@news.home.ibert.com> <73rflc$dku$5@strato.ultra.net> <1djdbbv.mlesnp1rxes0yN@n243-75.berlin.snafu.de> <749c8o$io4@srvr1.engin.umich.edu> <36690D71.4F611118@trailing-edge.com> <3677bac7.10795833@news.home.ibert.com> <36775b0a.17927949@news.ricochet.net> <1dk6q7m.fap2ctifdx6oN@n37-5.berlin.snafu.de> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On Fri, 18 Dec 1998 04:15:19 +0100, jnickelsen@acm.org (Juergen Nickelsen) wrote: >phil wrote: > >> A GUI is a bitch to develop for. Of all those available, NT is probably >> the easiest (you can produce a console app with minimal grief). > >I agree that console applications under Windows NT are fairly >straightforward (and can even be ported from Unix with little effort). >But this is not in any way related to *GUI* applications. I was actually talking about "console" applications - services to be precise (which are technically speaking not console applications when run the normal way because they don't really have a console). You _can_ port the UN*X code for "Hello, world!" rather easily, but if things get more complicated, the difference begin to make themselves felt. IPC and process control is totally different; the UN*X mechanisms either aren't there at all (fork/exec; signals), or they're quite different (pipes), or they just seem not to work too well sometimes, at least not if you use them as you would under UN*X (sockets), or they bite you when you least expect it by working differently (files, locking). If you structure a server process as you would under UN*X - okay, you can't, because there is no fork(). If you try to emulate that (in order to have common code), it will be very slow and de-stabilize the whole system if you use CreateProcess() or it will be somewhat slow and de-stabilize your application if you use threads (because now the serving instances share the address space). If you do it the way MS wants you to do it (with I/O completion ports and a pool of worker threads), your chances of retaining common source is minimal, because it is a major porting effort that is essentially one-way. I can't say anything about programming GUI applications, since I have no experience in that area. I hear that X is a bitch to develop for as well, but I can't judge that claim myself. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 05 Jan 1999 18:36:12 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 14 Message-ID: <36935b3c.188827145@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <74kofa$8tf@news1.newsguy.com> <74m8rf$5ev$1@vixen.cso.uiuc.edu> <3678c0d1.12342645@news.home.ibert.com> <757gi6$1504$1@nntp1.u.washington.edu> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!isdnet!howland.erols.net!news-feed.inet.tele.dk!bofh.vszbr.cz!newsfeed.tli.de!unlisys!news.snafu.de!home.ibert.com!news On 16 Dec 1998 05:30:46 GMT, dpeschel@u.washington.edu (D. Peschel) wrote: >Is there such a thing as an anti-bit? If so, Solaris contains the compiler >and development-library anti-bit most of the time, right? Compiler, yes. Development library, no. AFAIK, you can develop with Gnu C without having to buy the development system from Sun. (And Gnu C, for those who are not familiar with it, does not include the standard libraries or header files.) -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 05 Jan 1999 18:38:49 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 14 Message-ID: <36945c06.189028946@news.home.ibert.com> References: <73ela0$51h@hermes.acs.unt.edu> <366B7CFA.2A976248@tnglwood.demon.co.uk> <1998Dec7.174534.26644@lorelei.approve.se> <74hkig$oaq$1@news.wlg.netlink.net.nz> <74hp72$lk0$1@agate.berkeley.edu> <366f6b01.9489096@Rockyd> <3680c9be.14628000@news.home.ibert.com> <36780619.96636558@Rockyd> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newsfeed.uk.ibm.net!ibm.net!newsfeed.ecrc.net!unlisys!news.snafu.de!home.ibert.com!news On Wed, 16 Dec 1998 19:17:16 GMT, alex*@*rockvax.rockefeller.edu >clinfo> find / -name statfs -print >clinfo> > >Same with vstatfs. Any other ideas? Sorry, I didn't realize you wanted a command. statfs and vstatfs are system calls, with one of them usually being available on a UN*X system. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: martin@ibert.com (Martin Ibert) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Tue, 05 Jan 1999 18:44:02 GMT Organization: The Seventh Heaven, Berlin, Germany Lines: 21 Message-ID: <36955cb9.189207838@news.home.ibert.com> References: <749lip$gjf$1@vixen.cso.uiuc.edu> <74nqnu$1si$1@news.latrobe.edu.au> <1djtoy1.a1f9ypmi4mj7N@n241-84.berlin.snafu.de> <3684d146.16555558@news.home.ibert.com> <758rdo$12@artin.math.psu.edu> NNTP-Posting-Host: arezzo.home.ibert.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news-nyc.telia.net!masternews.telia.net!news-feed.inet.tele.dk!bofh.vszbr.cz!uninett.no!newsfeed.ecrc.net!unlisys!news.snafu.de!home.ibert.com!news On 16 Dec 1998 12:42:16 -0500, viro@artin.math.psu.edu (Alexander Viro) wrote: >Bzzzert. open("foo",O_RDONLY); unlink("foo"); and there you go - >file exists but is not accessible from the namespace. It will be >freed when the last process will close it. Link count is bloody >0. Ditto for binary of running process (at least on systems with >demand-loading). Ditto for a shared library. Link prevents file >from being freed, but there may be other holders. I know. It's just that I'd call this state on the disk "inconsistent", which may not the best word for it. (Sorry, not my native language.) If the filesystem is in this state when the machine comes up, fsck() will fix it. This state of the filesystem is only consistent if considered together with the in-core information of the operating system that is running - the filesystem itself, viewed in isolation, is inconsistent. -- >> Please visit http://www.ibert.com/ for contact information. << ----------------------------------------------------------------- Blind faith in your leaders, or in anything, will get you killed! ###### From: jnickelsen@acm.org (Juergen Nickelsen) Newsgroups: alt.folklore.computers Subject: Re: Is VMS taboo, or what? Date: Thu, 7 Jan 1999 20:17:09 +0100 Organization: [Posted via] Interactive Networx Lines: 18 Message-ID: <1dl8dns.130g9om7459woN@n35-14.berlin.snafu.de> References: <749lip$gjf$1@vixen.cso.uiuc.edu> <74nqnu$1si$1@news.latrobe.edu.au> <1djtoy1.a1f9ypmi4mj7N@n241-84.berlin.snafu.de> <3684d146.16555558@news.home.ibert.com> <758rdo$12@artin.math.psu.edu> <36955cb9.189207838@news.home.ibert.com> NNTP-Posting-Host: n35-14.berlin.snafu.de User-Agent: MacSOUP/2.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!howland.erols.net!news-feed.inet.tele.dk!bofh.vszbr.cz!newsfeed.tli.de!unlisys!news.snafu.de!jnickelsen Martin Ibert wrote: > This state of the filesystem is only consistent if considered together > with the in-core information of the operating system that is running - the > filesystem itself, viewed in isolation, is inconsistent. The first half of your sentence is very close to how I'd have worded it. With respect to the second half: considering the state of a mounted file system *without* the in-core information is rarely useful. Although, I have to add, one of the rare occasions is quite important -- when you think about what to do with this state when the disk comes back online after a power failure or similar event. But I think only file system (and file system checker) designers and implementors have to worry about that. -- Juergen Nickelsen