From: "T.W. Seddon" Newsgroups: alt.folklore.computers Subject: "640K should be enough for anyone" Date: 17 Feb 1998 18:23:09 GMT Organization: University of Newcastle upon Tyne Lines: 26 Message-ID: <6cckid$g10$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: mere1.ncl.ac.uk X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!feeder.qis.net!news-peer.sprintlink.net!news.sprintlink.net!Sprint!news-peer.gip.net!news-lond.gip.net!news.gsl.net!gip.net!peer.news.u-net.net!u-net!server5.netnews.ja.net!news.ncl.ac.uk!mere1.ncl.ac.uk!n5013784 I read often that Bill Gates was quoted as saying "640K should be enough for anyone", and that this is used against him. But, the amount of memory available to MS-DOS is fixed by the hardware, which was designed by IBM. The ROM and adapter space *must* live in the top X kilobytes, as the bottom of memory is used by the INT vector tables. As the decision was taken by IBM to reserve 384K for BIOS, adapter RAM, etc, Gates was stuck with the 640K RAM that was left. So, really, blame should be laid at the feet of IBM. Or should it? Did Intel not consider the possibility that with 20-bit addresses and vectors fixed at the bottom of memory the ROM etc would *have* to go at the top, limiting further expansion? Or did they think 1024K would be enough for anyone?! Maybe I'm taking Gates' quote out of context (to tell the truth, I don't know the exact context of the statement -- but given he was being paid by IBM and it was their hardware he could hardly say "640K is not enough for future expansion") but am I along the right lines? There are too many good reasons to villify (sp?) Bill Gates to waste time with the bad ones :-) -- --Tom ###### From: TheCentralScrutinizer.39@pobox.com () Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 17 Feb 1998 21:03:27 GMT Organization: Nyx Public Access Internet Lines: 8 Message-ID: References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> Reply-To: TheCentralScrutinizer.39@pobox.com NNTP-Posting-Host: iris.nyx.net X-Newsreader: slrn (0.9.4.3 UNIX) X-Disclaimer: Nyx is a Free Public Access Internet Service: http://www.nyx.net Our AUP / Free Speech Policy are at http://www.nyx.net/policies/ Direct complaints to abuse@nyx.net X-Post-Path: iris.nyx.net!afelson@nyx10.nyx.net Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsxfer3.itd.umich.edu!news-out.internetmci.com!newsfeed.internetmci.com!208.206.176.31!pulsar.dimensional.com!quasar.dimensional.com!nyx.net!TheCentralScrutinizer.39 In article <6cckid$g10$1@ucsnew1.ncl.ac.uk>, T.W. Seddon wrote: > >Or did they think 1024K would be enough for anyone?! they figured it was enough for an 8086/8088, and they were almost right. 1024K was a huge improvement compared to previous processors with only 16 bit addressing, only capable of addressing 64K. ###### From: ivie@cc.usu.edu (Roger Ivie) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Message-ID: Date: 17 Feb 98 21:20:07 MDT References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> Organization: Utah State University Lines: 16 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!atl-news-feed1.bbnplanet.com!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!newsfeed.internetmci.com!207.0.56.122!news.eli.net!inquo!xmission!news.cc.utah.edu!news.cs.utah.edu!cc.usu.edu!ivie In article <6cckid$g10$1@ucsnew1.ncl.ac.uk>, "T.W. Seddon" writes: > But, the amount of memory available to MS-DOS is fixed by the hardware, > which was designed by IBM. The ROM and adapter space *must* live in the > top X kilobytes, as the bottom of memory is used by the INT vector tables. But it doesn't have to _stay_ there. Z80s need ROM at location 0 for reset. CP/M needs RAM at location 0 for the base page. How did all those Z80 machines run CP/M? A popular trick was turning the ROM off during the boot sequence, replacing it with RAM. Could have been done on the PC. Could have used serial terminals, eliminating the need for memory mapped video. Both together would let you build an MS-DOS machine that can use an entire 1MB RAM. -- -------------------------+---------------------------------------------------- Roger Ivie | "You got advice for me. Well I'm telling you, ivie@cc.usu.edu | I know nothing and I like it that way" http://cc.usu.edu/~ivie/ | -- Eggplant ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 17 Feb 1998 22:53:42 GMT Message-ID: <6cd4dm$n7v$1@irk.zetnet.co.uk> References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: man-086.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Lines: 36 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!news-peer.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 1998-02-17 T.W.Seddon@ncl.ac.uk said: :Newsgroups: alt.folklore.computers :I read often that Bill Gates was quoted as saying "640K should be :enough for anyone", and that this is used against him. :But, the amount of memory available to MS-DOS is fixed by the :hardware, which was designed by IBM. The ROM and adapter space :*must* live in the top X kilobytes, as the bottom of memory is used :by the INT vector tables. As the decision was taken by IBM to :reserve 384K for BIOS, adapter RAM, etc, Gates was stuck with the :640K RAM that was left. yes. ms-dos wasn't limited to 640k; could potentially have used the whole megabyte if things had panned out that way. in the sirius 1 it could use up to 896k. :So, really, blame should be laid at the feet of IBM. Or should it? :Did Intel not consider the possibility that with 20-bit addresses :and vectors fixed at the bottom of memory the ROM etc would *have* :to go at the top, limiting further expansion? this is probably why the expansion they chose allowed remapping of daft physical positionings of memory to more logical (erm) logical positionings... ;> in any case, perhaps the problem is not so much the fact that the 8088 was limited to a megabyte, but *how* that megabyte was addressed. if only intel had made it switchable, so that either you could have the base*16+offset or an external mmu which did things properly... -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: hancock4@bbs.cpcn.com (Lisa or Jeff) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 18 Feb 1998 01:08:52 GMT Organization: Net Access BBS Lines: 16 Message-ID: <6cdcb4$9oc@netaxs.com> NNTP-Posting-Host: bbs.cpcn.com Originator: root@bbs.cpcn.com Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsxfer3.itd.umich.edu!news-peer.gip.net!news.gsl.net!gip.net!news-xfer.netaxs.com!netaxs.com!bbs.cpcn.com!root > not consider the possibility that with 20-bit addresses and vectors fixed > at the bottom of memory the ROM etc would *have* to go at the top, limiting > further expansion? > Or did they think 1024K would be enough for anyone?! It must be remembered that the technology of those days was quite different than today. 15 years is a long time, especially in electronics. Windows didn't exist yet. 32Meg of RAM on a desktop would have been big and bulky and extremely expensive. As someone else mentioned, this configuration was a big improvement over other microcomputer technologies. Don't forget the original IBM PC had maxed out at 640k, as an extra cost option. Smaller memories were available. ###### From: w6nh2@jupiter.sun.csd.unb.ca (Sir Isle) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 18 Feb 1998 04:33:42 GMT Organization: University of New Brunswick, Fredericton, NB, Canada Lines: 11 Message-ID: <6cdob6$r9b@sol.sun.csd.unb.ca> References: <6cdcb4$9oc@netaxs.com> NNTP-Posting-Host: sol-alt1.unb.ca X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!xfer.kren.nm.kr!news-peer.gip.net!news.gsl.net!gip.net!europa.clark.net!199.60.229.5!newsfeed.direct.ca!torn!garnet.nbnet.nb.ca!news.unb.ca!sol!w6nh2 On 18 Feb 1998 Lisa or Jeff (hancock4@bbs.cpcn.com) uttered forth the following: : It must be remembered that the technology of those days was quite different : than today. 15 years is a long time, especially in electronics. Windows : didn't exist yet. 32Meg of RAM on a desktop would have been big and : bulky and extremely expensive. In those days, 32Meg of hard drive space was a lot, considering the largest XT hard drive I've seen was 20Meg. I once picked up an old XT with a full height 8Meg hard drive in it; I was tempted to keep the drive just as a souvenir. :) Isle ###### From: R!ch Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Wed, 18 Feb 1998 07:45:20 +0000 Lines: 39 Message-ID: References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: paddington.uk.sun.com Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Sender: richardt@paddington To: "T.W. Seddon" In-Reply-To: <6cckid$g10$1@ucsnew1.ncl.ac.uk> Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsxfer3.itd.umich.edu!news-out.internetmci.com!newsfeed.internetmci.com!194.72.7.126!btnet-peer!btnet-feed2!btnet!carbon.eu.sun.com!uk-usenet.uk.sun.com!paddington!richardt On 17 Feb 1998, T.W. Seddon wrote: > But, the amount of memory available to MS-DOS is fixed by the hardware, > which was designed by IBM. The ROM and adapter space *must* live in the > top X kilobytes, as the bottom of memory is used by the INT vector tables. > As the decision was taken by IBM to reserve 384K for BIOS, adapter RAM, etc, > Gates was stuck with the 640K RAM that was left. This is true to a degree. However, there's no reason why M$ couldn't take advantage of the 386's MMU to implement some form of VM. That way, the hole in the middle would be a non-issue. > So, really, blame should be laid at the feet of IBM. Or should it? Did Intel > not consider the possibility that with 20-bit addresses and vectors fixed > at the bottom of memory the ROM etc would *have* to go at the top, limiting > further expansion? I personally think that IBM screwed up when they based the PC on the 8088/6 - I think they should have gone for the 68000. If they had done so, the peecee step to 32 bits would have been a lot easier (not that it affected me - I've never owned one). -- R!ch (Email is flakey at present: use richardt@keaton.uk.sun.com) If it ain't analogue, it ain't music. #include \\|// - ? (o o) /==================================oOOo=(_)=oOOo========\ | Richard Teer richard.teer@uk.sun.com | | | | | | WWW: www.rkdltd.demon.co.uk | | .oooO | | ( ) Oooo. | \===================================\ (==( )==========/ \_) ) / (_/ ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 18 Feb 1998 20:57:47 GMT Message-ID: <6cfi0b$q9o$4@irk.zetnet.co.uk> References: NNTP-Posting-Host: man-171.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Lines: 21 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!woodstock.news.demon.net!demon!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-18 6cckid$g10$1@ucsnew1.ncl.ac.uk said: :I personally think that IBM screwed up when they based the PC on the :8088/6 - I think they should have gone for the 68000. If they had :done so, the peecee step to 32 bits would have been a lot easier :(not that it affected me - I've never owned one). deja vu... :> besides, knowing microsoft, they would have used the top 8 bits for something interesting inside dos; knowing ibm, they still would have used the top 8 bits inside the bios; and we still would have been stuck with a single-user, single-tasking program loader struggling to support a dangerously top-heavy kludged together gui which pretends to be an os. -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: "Ralph Wade Phillips" Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 19 Feb 1998 02:37:09 GMT Organization: Phillips Enterprises Lines: 17 Message-ID: <01bd3ce0$6b257640$f62cd6d0@ralphp.shreve.net> References: <6cdcb4$9oc@netaxs.com> NNTP-Posting-Host: 208.214.44.246 X-Newsreader: Microsoft Internet News 4.70.1161 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!news.vt.edu!solaris.cc.vt.edu!news.cc.ukans.edu!news.starnet.net!sdd.hp.com!vixen.cso.uiuc.edu!logbridge.uoregon.edu!news.shreve.net!not-for-mail Hi, Lisa or Jeff! Lisa or Jeff wrote in article <6cdcb4$9oc@netaxs.com>... (stuff deleted for bandwidth preservation) > > Don't forget the original IBM PC had maxed out at 640k, as an extra cost > option. Smaller memories were available. > > Errm ... Due to a bug in the BIOS, the ORIGINAL PC maxed out at 576K ... RwP ###### From: Ian Stirling <000034ED3FED.NO_UCE@mauve.demon.co.uk> Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Fri, 20 Feb 1998 08:33:49 GMT Organization: None. Message-ID: <000034ED3FED.NO_UCE@mauve.demon.co.uk> References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> <01bd3dcb$10752510$2fba36cf@alpha> X-Mail2News-User: Send.NO_UCE@mauve.demon.co.uk X-Mail2News-Path: post-10.mail.demon.net!post.mail.demon.net!mauve.demon.co.uk X-Trace: mail2news.demon.co.uk 888033238 3711 Send.NO_UCE mauve.demon.co.uk X-Complaints-To: abuse@demon.net User-Agent: tin/pre-1.4-971106 (UNIX) (Linux/2.0.33 (i486)) X-Note: Anti-UCE in effect, replying should work if you are not UCE'ng X-Warning0: For unsolicited commercial email, sent or causing to be sent to my email address X-Warning1: on this message, I reserve the right to levy a charge for my time and expenses X-Warning2: of up to 100 pounds sterling per message, plus legal, penalty or other costs. Lines: 25 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!woodstock.news.demon.net!demon!news.demon.co.uk!demon!mail2news.demon.co.uk!mauve.demon.co.uk!000034ED3FED.NO_UCE Eugene A. Pallat wrote: : T.W. Seddon wrote in article : <6cckid$g10$1@ucsnew1.ncl.ac.uk>... :> I read often that Bill Gates was quoted as saying "640K should be enough :> for anyone", and that this is used against him. :> :> But, the amount of memory available to MS-DOS is fixed by the hardware, :> which was designed by IBM. The ROM and adapter space *must* live in the :> top X kilobytes, as the bottom of memory is used by the INT vector : tables. : No - that's a totally false popular assumption. The int vector addresses : can be software mapped to go anywhere, such as a transfer vector. ROM can : also go anywhere, but the low addresses are best. That way you don't have Surely, at the time the original PC was designed, assuming that they went with the 8086/88, does this not mandate haveing the int vectors in low RAM? I admit it's been a while. -- Ian Stirling. Designing a linux PDA, see http://www.mauve.demon.co.uk/ ----- ******* If replying by email, check notices in header ******* ----- Money is a powerful aphrodisiac, but flowers work almost as well. Robert A Heinlein. ###### From: Marco S Hyman Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 20 Feb 1998 12:46:15 -0800 Organization: S.N.A.F.U. (www.snafu.org) Lines: 14 Message-ID: References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> <01bd3dcb$10752510$2fba36cf@alpha> <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk> <6ckfn6$hg6$1@eagle.eku.edu> NNTP-Posting-Host: dumbcat.codewright.com X-Trace: 888007582 15818 marc 206.86.0.12 X-Newsreader: Gnus v5.5/Emacs 20.2 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news-peer.gip.net!news.gsl.net!gip.net!news-peer.sprintlink.net!news.sprintlink.net!Sprint!newsxfer3.itd.umich.edu!news1.best.com!nntp2.ba.best.com!not-for-mail styer@eagle.eku.edu (Gene Styer) writes: > Actually we have another 8086 problem - The chip was wired to start executing > instructions at FFFF:0000, so it would have been difficult to put ROM anywhere > else. Yes. But I've worked with several embedded devices where the rom did what it had to do, including reloate any part of itself that needed to be around at run time, and then jump to the relocated code. The very first thing the relocated code did was map the rom out and ram in. This would often be something like the top 256 or 512k. A hard reset would do the opposite. // marc ###### Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Message-ID: <6ckfn6$hg6$1@eagle.eku.edu> From: styer@eagle.eku.edu (Gene Styer) Date: 20 Feb 1998 12:49:26 -0500 References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> <01bd3dcb$10752510$2fba36cf@alpha> <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk> Organization: Eastern Kentucky University NNTP-Posting-Host: eagle.eku.edu Lines: 40 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!feed2.news.erols.com!erols!news-peer.sprintlink.net!news.sprintlink.net!Sprint!newsfeed.internetmci.com!204.71.76.137!news.campus.mci.net!uky.edu!acs.eku.edu!not-for-mail In article <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk>, T.W. Seddon wrote: >Eugene A. Pallat (eapallat@orion-data.com) wrote: >> T.W. Seddon wrote: >> > But, the amount of memory available to MS-DOS is fixed by the hardware, >> > which was designed by IBM. The ROM and adapter space *must* live in the >> > top X kilobytes, as the bottom of memory is used by the INT vector >> tables. > >> No - that's a totally false popular assumption. The int vector addresses >> can be software mapped to go anywhere, such as a transfer vector. ROM can >> also go anywhere, but the low addresses are best. That way you don't have >> a ROM address blocking the address range That's the way we designed our >> embedded control trainer system to work. > >How can this be done? I see no register on the 8086 that specifies the >base address of the INT vectors -- the early Intel chips had the vectors >fixed at 0000:0000 if I remember right. The 68000 was the same. Yep - the 8086 had this fixed at 0000:0000. Remember all of the protected mode stuff (including the Interrupt Description Table) was introduced in the 286, and extended in the 386. I did a 68000 based project during graduate school, and the 68000 had the restart vector (which needed to be rom) at 0, and various interrupt vectors (which generally we wanted to be ram) at 4, 8, 12, etc. This made for some interesting problems. One solution was to have the address decoder always to go ROM for the first N cycles, and then switch to normal decoding after the key words had been fetched. As I think about it, we might have waited until it started executing instructions instead of counting cycles, but the approach was the same. > >I suppose you could have ROM just above the vector tables, but this strikes >me as a little odd. But then, if it would have solved the 640K problem... Actually we have another 8086 problem - The chip was wired to start executing instructions at FFFF:0000, so it would have been difficult to put ROM anywhere else. ###### From: hotlynx@SPAMBEGONEwhidbey.net Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Fri, 20 Feb 1998 13:45:19 -0800 Organization: WhidbeyNet News Service Lines: 19 Message-ID: References: <6cdcb4$9oc@netaxs.com> <6cdob6$r9b@sol.sun.csd.unb.ca> NNTP-Posting-Host: asn75.whidbey.net X-Newsreader: Anawave Gravity v2.00 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!news-peer.sprintlink.net!news-backup-west.sprintlink.net!news-in-west.sprintlink.net!news.sprintlink.net!Sprint!204.94.52.5!news.whidbey.com!not-for-mail In article <6cdob6$r9b@sol.sun.csd.unb.ca>, w6nh2@jupiter.sun.csd.unb.ca says... > On 18 Feb 1998 Lisa or Jeff (hancock4@bbs.cpcn.com) uttered forth the following: > : It must be remembered that the technology of those days was quite different > : than today. 15 years is a long time, especially in electronics. Windows > : didn't exist yet. 32Meg of RAM on a desktop would have been big and > : bulky and extremely expensive. > > In those days, 32Meg of hard drive space was a lot, considering the > largest XT hard drive I've seen was 20Meg. I once picked up an old XT with a > full height 8Meg hard drive in it; I was tempted to keep the drive just as > a souvenir. :) > Isle > I had a 40 MB drive on an XT, but it required partitioning because 32 MB was all the BIOS could handle. -- L. Nino -- You can blame everything on me this year. ###### From: "T.W. Seddon" Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 20 Feb 1998 16:02:15 GMT Organization: University of Newcastle upon Tyne Lines: 34 Message-ID: <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk> References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> <01bd3dcb$10752510$2fba36cf@alpha> NNTP-Posting-Host: pike9.ncl.ac.uk X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!nntp.news.xara.net!xara.net!server6.netnews.ja.net!news.ncl.ac.uk!pike9.ncl.ac.uk!n5013784 Eugene A. Pallat (eapallat@orion-data.com) wrote: > T.W. Seddon wrote: > > But, the amount of memory available to MS-DOS is fixed by the hardware, > > which was designed by IBM. The ROM and adapter space *must* live in the > > top X kilobytes, as the bottom of memory is used by the INT vector > tables. > No - that's a totally false popular assumption. The int vector addresses > can be software mapped to go anywhere, such as a transfer vector. ROM can > also go anywhere, but the low addresses are best. That way you don't have > a ROM address blocking the address range That's the way we designed our > embedded control trainer system to work. How can this be done? I see no register on the 8086 that specifies the base address of the INT vectors -- the early Intel chips had the vectors fixed at 0000:0000 if I remember right. The 68000 was the same. I suppose you could have ROM just above the vector tables, but this strikes me as a little odd. But then, if it would have solved the 640K problem... Actually, whilst looking for something *completely* different the other night, I found on one of my shareware CDs a copy of the alt.folklore. computers FAQ from 1992 (I didn't even know a FAQ existed!). This claimed that IBM realised 1024K was a hopeless amount of memory but they were intending to do an IBM PC mk II, which would eliminate many of the problems -- although as most of the problems stem from that b******d 80x86 architecture it's hard to see how they could have fixed it without breaking existing software. The story goes on to state what we now all know: that this IBM PC mk II never materialised. The rest, as they say, is history ;-) --Tom ###### From: cbh@REMOVE_THIS.teabag.demon.co.uk (Chris Hedley) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 20 Feb 1998 19:41:58 GMT Organization: teabag Message-ID: <6ckma6$10a$1@teabag.demon.co.uk> References: <6ci7i5$ta3$4@irk.zetnet.co.uk> NNTP-Posting-Host: localhost X-NNTP-Posting-Host: teabag.demon.co.uk [193.237.4.110] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Newsreader: knews 0.9.8 Lines: 23 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!woodstock.news.demon.net!demon!news.demon.co.uk!demon!teabag.demon.co.uk!not-for-mail In article <6ci7i5$ta3$4@irk.zetnet.co.uk>, lisard@zetnet.co.uk writes: > some msdos machines (some were even ibm-compatible...) did use serial > terminals, but we're glad that it didn't catch on... memory mapping may > have eaten ram, but it does give you the maximum amount of control over > what appears on your monitor. Oh, I dunno; packages like Unix' curses and VAXforms (or whatever it was called) allowed just as much control over what appeared on the screen as a memory-mapped character based display; the only downside is that it'd typically be a bit slower (perhaps), but not unusably slow. I'm still quite a fan of the "old fashioned" green-screen terminals, and some of the newer ones run at a high enough speed to give almost instantaneous response which, coupled with a multi-session adapter (hardware or software based, the latter being preferable as it does automatic screen refreshes, even with applications that wouldn't normally), actually provide quite a nice user environment. I know that all this may sound a little out of date to some people but I still maintain that a properly thought- out VT style user environment beats the hell out of many, perhaps even most, windows-based rivals (okay, I know I'm using the X11 based Knews to read this newsgroup, but I'm still using vi-in-an-xterm to reply. :) Chris. ###### From: "Greg Menke" Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 20 Feb 1998 20:21:44 GMT Organization: .. Lines: 75 Message-ID: <01bd3e3d$1b0520d0$64646464@gregm> References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> <01bd3dcb$10752510$2fba36cf@alpha> <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: 207-172-155-149.s70.as19.col.erols.com X-Trace: winter.news.erols.com 888006104 23885 207.172.155.149 (20 Feb 1998 20:21:44 GMT) X-Complaints-To: abuse@erols.com X-Newsreader: Microsoft Internet News 4.70.1162 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!news.ecn.uoknor.edu!feed1.news.erols.com!winter.news.erols.com!not-for-mail I've never heard of being able to change the location of the interrupt table, at least I'm certain you can't on the 80186/8 or 8086/8. Although I've never used protected mode, I imagine you would assign an interrupt table for each task on processors >= 80286 - sort of achieving the interrupt table relocation. As I recall, all x86 processors (not sure of the Pentium), start in real mode, beginning execution at a word or so below the top of memory- forcing ROM startup code to be located there. Of course, once the user bios gains control it can do whatever it wants to the memory map. Real mode also defines the interrupt table to be in low memory. This does impose a specific startup model to the memory architecture, but once started, its all up to the user. In real mode, the amount of linearly addressable memory does have a fixed upper limit (1 meg), simply because the chip won't generate the extended addresses (IIRC). As far as memory mapped devices are concerned, the conflict with ROM space is THE problem, and thats why you design the memory architecture to avoid it. And thats also the reason you have all QEMM/QRAM/EMM386/396MAX and friends on the PC - to switch addressing modes and page in and out the "extended memory" above 1 meg. IO mapped devices don't suffer this particular trouble, the space is more limited (16 bits) and you still have to partition it usefully. Given the segmented architecture its pretty easy to get confused with what linear addresses are being generated from the segment:offsets. -- To reply, replace nospam with menkesjg. Thanks. T.W. Seddon wrote in article <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk>... > Eugene A. Pallat (eapallat@orion-data.com) wrote: > > T.W. Seddon wrote: > > > But, the amount of memory available to MS-DOS is fixed by the hardware, > > > which was designed by IBM. The ROM and adapter space *must* live in the > > > top X kilobytes, as the bottom of memory is used by the INT vector > > tables. > > > No - that's a totally false popular assumption. The int vector addresses > > can be software mapped to go anywhere, such as a transfer vector. ROM can > > also go anywhere, but the low addresses are best. That way you don't have > > a ROM address blocking the address range That's the way we designed our > > embedded control trainer system to work. > > How can this be done? I see no register on the 8086 that specifies the > base address of the INT vectors -- the early Intel chips had the vectors > fixed at 0000:0000 if I remember right. The 68000 was the same. > > I suppose you could have ROM just above the vector tables, but this strikes > me as a little odd. But then, if it would have solved the 640K problem... > > Actually, whilst looking for something *completely* different the other > night, I found on one of my shareware CDs a copy of the alt.folklore. > computers FAQ from 1992 (I didn't even know a FAQ existed!). This > claimed that IBM realised 1024K was a hopeless amount of memory but they > were intending to do an IBM PC mk II, which would eliminate many of the > problems -- although as most of the problems stem from that b******d > 80x86 architecture it's hard to see how they could have fixed it without > breaking existing software. > > The story goes on to state what we now all know: that this IBM PC mk II > never materialised. The rest, as they say, is history ;-) > > --Tom > > ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 20 Feb 1998 21:58:51 GMT Message-ID: <6ckuar$eq$6@irk.zetnet.co.uk> References: <6ck9e7$n6p$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: man-093.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Lines: 23 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!woodstock.news.demon.net!demon!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-20 T.W.Seddon@ncl.ac.uk said: :How can this be done? I see no register on the 8086 that specifies :the base address of the INT vectors -- the early Intel chips had :the vectors fixed at 0000:0000 if I remember right. The 68000 was :the same. indeed. the 286 could move them (??) but not the 8086. :I suppose you could have ROM just above the vector tables, but this :strikes me as a little odd. But then, if it would have solved the :640K problem... well, you could have rom for the vector tables, and just redirect them to a jump table which then jumped on to the real handler, and would live somewhere completely different in ram... -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 21 Feb 1998 12:57:01 GMT Lines: 49 Message-ID: <6cmiut$2dt$1@irk.zetnet.co.uk> References: <6ckma6$10a$1@teabag.demon.co.uk> NNTP-Posting-Host: man-138.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!feeder.qis.net!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-20 cbh@REMOVE_THIS.teabag.demon.co.uk(ChrisHedley) said: :Oh, I dunno; packages like Unix' curses and VAXforms (or whatever :it was called) allowed just as much control over what appeared on :the screen as a memory-mapped character based display; the only :downside is that it'd typically be a bit slower (perhaps), but not :unusably slow. ...provided you have a high enough speed on the serial link. however, can you tell the terminal "you just echo everything you get in to the screen, and i'll read it from there, thank you"? that's the sort of trick we're thinking of pulling with a forth system, to get rid of the necessity to treat the keyboard and screen as a terminal. also, it seems to be what a lot of home computers did - remember the commodore full screen editors? :I'm still quite a fan of the "old fashioned" :green-screen terminals, and some of the newer ones run at a high :enough speed to give almost instantaneous response which, coupled :with a multi-session adapter (hardware or software based, the :latter being preferable as it does automatic screen refreshes, even :with applications that wouldn't normally), actually provide quite a :nice user environment. hmm. we don't mind the idea of terminals at all, except that you tend to be stuck with not very many lines on the screen and fixed-width fonts. (in our view the only good thing to come out of guis are proportional fonts...) but if you want to keep your cpu for doing cpu-y things, or want a multi-uiser box, vts are probably the way to go. (especially if you want to debug on one console and run on another. sharing a keyboard between 2 virtual consoles isn't always nice...) :I know that all this may sound a little out :of date to some people but I still maintain that a properly :thought- out VT style user environment beats the hell out of many, :perhaps even most, windows-based rivals (okay, I know I'm using the :X11 based Knews to read this newsgroup, but I'm still using :vi-in-an-xterm to reply. :) but then a well-thought-out command line system beats the hell out of a gui too. :> and we're using dos software for internet access, having singularly failed to set up anything else... -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: dfox@belvdere.vip.best.com (David E. Fox) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 21 Feb 1998 13:19:08 -0800 Organization: Dave's Really K-rad Linux Box Lines: 21 Message-ID: References: <6cckid$g10$1@ucsnew1.ncl.ac.uk> Reply-To: dfox@belvdere.vip.best.com NNTP-Posting-Host: belvdere.vip.best.com X-Trace: 888096191 23079 dfox 206.86.0.12 X-Newsreader: slrn (0.9.4.2 UNIX) Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!news1.best.com!nntp2.ba.best.com!dfox In article , TheCentralScrutinizer.39@pobox.com wrote: >they figured it was enough for an 8086/8088, and they were almost right. I think for a time, they were right. It was only really when corporations started adapting 1-2-3 (the real killer PC app at the time) and started to put their huge spreadsheets on it when the 640K limit became readily apparent. Before then it was *a lot* of memory. I remember the day I got my XT clone with 640K RAM and I was so thrilled - a _tenfold_ increase in RAM overnight (my previous system was a 64K level 1 trs80). Now, I've got 48 megs in a Linux box :). -- ------------------------------------------------------------------------ David E. Fox Tax Thanks for letting me dfox@belvdere.vip.best.com the change magnetic patterns root@belvedere.sbay.org churches on your hard disk. ----------------------------------------------------------------------- ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 21 Feb 1998 16:32:16 GMT Lines: 30 Message-ID: <6cmvig$2q4$1@irk.zetnet.co.uk> References: <6cfkc5$8oe$1@ucsnew1.ncl.ac.uk> NNTP-Posting-Host: man-068.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news-peer.gip.net!news-lond.gip.net!news.gsl.net!gip.net!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-18 T.W.Seddon@ncl.ac.uk said: :> I personally think that IBM screwed up when they based the PC on :>the 8088/6 - I think they should have gone for the 68000. If :>they had done so, the peecee step to 32 bits would have been a :>lot easier (not that it affected me - I've never owned one). :I've always thought the 680x0 series would have been far better :than the nasty old x86 series. But, the 68000's address bus was :limited to 24 bits, and the TRAP vectors are at the bottom of :memory, so I suppose we'd all be complaining about the hopeless :CP/M-68K 15MB barrier :-) How are we supposed to get decent :performance out of GEM '95 in that amount of memory then? :-) nah, performance wouldn't be a problem, we'd all be moaning about how to get decent fonts and overlapping windows out of it. :> (but looking over our shoulder at that now-defunct sad little company called microsoft who at one time threatened to take over the world, and thanking all the powers that be that the crappy windows effort never took off...) on the other hand - cp/m-68k? the do-it-yourself operating system? :> we always got the impression that dr were reluctant to support a chip that didn't say intel on it. (at least until jack tramiel had a word.) -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: Ben Hutchings Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 21 Feb 1998 22:06:06 -0000 Organization: Not organised Lines: 39 Sender: womble@eldritch.dyn.ml.org Message-ID: <6cnj4e$3db$1@eldritch.dyn.ml.org> References: <6cdcb4$9oc@netaxs.com> <6cdob6$r9b@sol.sun.csd.unb.ca> NNTP-Posting-Host: max82.public.ox.ac.uk Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!nntp-out.monmouth.com!newspeer.monmouth.com!logbridge.uoregon.edu!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!server5.netnews.ja.net!server3.netnews.ja.net!news.ox.ac.uk!not-for-mail In article , wrote: >In article <6cdob6$r9b@sol.sun.csd.unb.ca>, w6nh2@jupiter.sun.csd.unb.ca >says... >> On 18 Feb 1998 Lisa or Jeff (hancock4@bbs.cpcn.com) uttered forth the following: >> : It must be remembered that the technology of those days was quite different >> : than today. 15 years is a long time, especially in electronics. Windows >> : didn't exist yet. 32Meg of RAM on a desktop would have been big and >> : bulky and extremely expensive. >> >> In those days, 32Meg of hard drive space was a lot, considering the >> largest XT hard drive I've seen was 20Meg. I once picked up an old XT with a >> full height 8Meg hard drive in it; I was tempted to keep the drive just as >> a souvenir. :) >> Isle >> >I had a 40 MB drive on an XT, but it required partitioning because 32 MB >was all the BIOS could handle. No, that's wrong. Partitioning has nothing to do with the BIOS. The 32 MB limitation was in DOS. The FAT16 file-system used for hard-disks had 16-bit block numbers, and each block stored 512 bytes. Therefore it could only address up to 32 MB. (The original FAT12 file-system used for floppies could only go up to 2 MB, naturally.) From DOS 4.0 onwards the blocks were expanded into "clusters" that could be 512*2^n bytes, for some integer n. By the time 2 GB drives were around, of course, the clusters had grown to 32 KB, making the majority of the drives' contents wasted space (not that you'd expect any different from Windows softare). It was worth using DoubleSpace with compression disabled just to get round this, as it uses an alternative file-system without this silly allocation scheme. The newer FAT32 file-system used in later versions of Windows 95 and in Windows 98 uses 32-bit numbers for allocation units and so can be used to divide even the largest drives up into 512-byte units. -- Ben Hutchings, M&CS student | Jay Miner Society website: http://www.jms.org/ email/finger m95bwh@ecs.ox.ac.uk | homepage http://users.ox.ac.uk/~worc0223/ Q: How can you tell when a lawyer is lying? A: His lips are moving. ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 22 Feb 1998 00:28:54 GMT Message-ID: <6cnrg6$3us$1@irk.zetnet.co.uk> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> NNTP-Posting-Host: man-188.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Lines: 47 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news-peer.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 1998-02-21 worc0223@sable.ox.ac.uk said: :>I had a 40 MB drive on an XT, but it required partitioning because :>32 MB was all the BIOS could handle. :No, that's wrong. Partitioning has nothing to do with the BIOS. :The 32 MB limitation was in DOS. The FAT16 file-system used for :hard-disks had 16-bit block numbers, and each block stored 512 :bytes. Therefore it could only address up to 32 MB. (The original :FAT12 file-system used for floppies could only go up to 2 MB, :naturally.) closer, but still no cigar. the 32Mb limitation was in dos, but wasn't anything to do with the fat (which always worked in clusters; dos 1 formatted disks so that each fat entry, or cluster, referred to 2 sectors; that allowed it to fit into a single disk sector.) it's to do with int 0x25 and 0x26, which until dos 4 would only accept 16-bit absolute sector numbers. the bios could always cope with larger drives, since it used a cylinder/sector/head numbering scheme that only finally ran out of steam when drives larger than 4Gb came along. (how *do* you access them, please, someone?) fat12 could cope quite happily with drives larger than 2Mb; indeed, if you pushed hard enough it could go right up to 32Mb - but at a cost of using 8k clusters, which was felt to lead to too much fragmentation, so dos 3 introduced fat16 and used 2k clusters, but only if you asked it to deal with a partition larger than 16Mb. under that, it went happily back to fat12, and for all we know the latest versions of dos still do. (anybody fancy creating a 12Mb partition and testing this theory...? oh, hang on, we'll do it ourselves.) the *real* problem with the fat is that it's basically a linear access system; to get to the last cluster in your file, you have to read all the fat entries in. great, if they're all together; tough, if they're scattered across the length of your hard drive. there have been some much more sensible schemes for sector allocation conceived, our personal favourite being the scheme used in thoth (to get back to folklore :> ). btw - is thoth online anywhere? we have the book, we want the source. :> -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: "Jeffrey S. Dutky" Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Sun, 22 Feb 1998 01:02:26 -0500 Organization: University of Maryland Lines: 33 Message-ID: <34EFBF68.5707@bellatlantic.net> References: <6cfi0b$q9o$4@irk.zetnet.co.uk> <6chb3e$vod$1@nnrp1.dejanews.com> Reply-To: dutky@bellatlantic.net NNTP-Posting-Host: client-151-200-122-29.bellatlantic.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.04Gold (Macintosh; I; PPC) Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!nntp-out.monmouth.com!newspeer.monmouth.com!howland.erols.net!world6.bellatlantic.net!news jk@langley.softwright.co.uk wrote: > Well, those spare 8 bits certainly tempted Apple and a > smattering of third parties who got burnt when the System > moved from 24 to 32-bit addressing -- as an SE/30 user I > remember the Mode32 patch to fix up its dirty ROMs. > > Does anyone remember if any of the major apps got hit on > 32-bit compatibility? None spring to mind right now, but > I've been away from the Mac world for several years... > Not that I actually REMEMBER any specific apps that got hit by the 32-bit clean problem, but I'm certain that a fair number actually DID break on that one. The reason noone much noticed is two-fold: 1) Most Macs couldn't use 32-bit addressing at the time. The majority of the Mac installed base were Pluses and SEs, while a relative minority of systems were II's or SE/30s. In this circumstance the number of users ABLE to detect 32-bit dirty software was fairly low. 2) Apple made a BIG deal about the comming 32-bit clean addressing problem and most software was released in a fixed version shortly after the MacOS required it. If you want to know what software wasn't 32-bit clean at the time you could check out any old lists of AU/X complatible applications. Remember that AU/X was able to run MacOS apps in a compatability box setup, but only if the apps were 32-bit clean. - Jeff Dutky ###### From: J. Chris Hausler Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Sun, 22 Feb 98 15:13:43 -0500 Organization: Delphi (info@delphi.com email, 800-695-4005 voice) Lines: 25 Message-ID: References: <6ckma6$10a$1@teabag.demon.co.uk> <6cmiut$2dt$1@irk.zetnet.co.uk> NNTP-Posting-Host: 199.93.4.4 X-To: Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!vixen.cso.uiuc.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!cam-news-feed1.bbnplanet.com!news.bbnplanet.com!news.delphi.com!news writes: >...provided you have a high enough speed on the serial link. however, >can you tell the terminal "you just echo everything you get in to the >screen, and i'll read it from there, thank you"? that's the sort of >trick we're thinking of pulling with a forth system, to get rid of the >necessity to treat the keyboard and screen as a terminal. also, it seems >to be what a lot of home computers did - remember the commodore full >screen editors? My last 8 bit system, a SWTPC 6809, had a full screen editor, named Stylograph, which wasn't too bad. Every time you hit a key to enter text, the system would start rewriting the screen from the point of the new character until either the end of the screen or, more likely, until the added character no longer affected the screen layout. At 9600 bps, this normally worked fine. If however, you had a full screen of text and started adding characters at the top of the screen and typed quickly you could get ahead of it. Every time you hit a new key, if it wasn't done updating the screen, it would quit and start over at the new character you had just entered. Eventually you would have gobblygook on the lower part of the screen. Of course, as soon as you paused for more than a second or so, it would complete the update. I could never keep this going for more than a few seconds but it was fun in a strange way :-) Chris ###### From: Ben Hutchings Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 22 Feb 1998 17:27:14 -0000 Organization: Not organised Lines: 20 Sender: womble@eldritch.dyn.ml.org Message-ID: <6cpn5i$5i3$1@eldritch.dyn.ml.org> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> NNTP-Posting-Host: max87.public.ox.ac.uk Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news-peer.gip.net!news.gsl.net!gip.net!news-peer.sprintlink.net!news.sprintlink.net!Sprint!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!rill.news.pipex.net!pipex!server1.netnews.ja.net!server2.netnews.ja.net!news.ox.ac.uk!not-for-mail In article <6cnrg6$3us$1@irk.zetnet.co.uk>, wrote: >the bios could always cope with larger drives, since it used a >cylinder/sector/head numbering scheme that only finally ran out of >steam when drives larger than 4Gb came along. (how *do* you access >them, please, someone?) Actually, I thought the problem cropped up at 2 GB. Anyway, it is fixed by hitting the disk controller directly. The limits on C/H/S dimensions in the ATA spec are different from those in the BIOS settings, so there was an earlier problem at around the 504 MB mark. This was fixed by the introduction of logical block addressing, which SCSI drives have been using since, er, quite a while back. -- Ben Hutchings, M&CS student | Jay Miner Society website: http://www.jms.org/ email/finger m95bwh@ecs.ox.ac.uk | homepage http://users.ox.ac.uk/~worc0223/ Adding manpower to a late software project makes it later. - Fred Brooks ###### From: cbh@REMOVE_THIS.teabag.demon.co.uk (Chris Hedley) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 22 Feb 1998 18:57:13 GMT Organization: teabag Message-ID: <6cpse9$4jo$2@teabag.demon.co.uk> References: <6ckma6$10a$1@teabag.demon.co.uk> <6cmiut$2dt$1@irk.zetnet.co.uk> NNTP-Posting-Host: localhost X-NNTP-Posting-Host: teabag.demon.co.uk [193.237.4.110] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Newsreader: knews 0.9.8 Lines: 68 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!woodstock.news.demon.net!demon!news.demon.co.uk!demon!teabag.demon.co.uk!not-for-mail In article <6cmiut$2dt$1@irk.zetnet.co.uk>, lisard@zetnet.co.uk writes: > ..provided you have a high enough speed on the serial link. however, > can you tell the terminal "you just echo everything you get in to the > screen, and i'll read it from there, thank you"? that's the sort of > trick we're thinking of pulling with a forth system, to get rid of the > necessity to treat the keyboard and screen as a terminal. also, it seems > to be what a lot of home computers did - remember the commodore full > screen editors? Well, that's why I mentioned the curses and forms packages, which solve both the bandwidth and the "read back" problems. They keep a copy of what's on display & the relevant attributes in memory, so if you need to reference something it's dead easy, and to keep bandwidth down they figure out how to send the absolute minimum of data across the serial line when an update is required. Works quite nicely even at comparitively low line speeds; my experience is generally with screens at 9600 baud, and it is quite acceptable at much lower speeds, although full-screen redraws can be slow at very low speeds (like 1200bps and lower) Modern dumb terminals run at much higher speeds, in the region of 110,000bps if my memory serves (okay, I know that 3270 terminals have run at line speeds in the megabit range for decades now!) > hmm. we don't mind the idea of terminals at all, except that you tend to > be stuck with not very many lines on the screen and fixed-width fonts. > (in our view the only good thing to come out of guis are proportional > fonts...) but if you want to keep your cpu for doing cpu-y things, or > want a multi-uiser box, vts are probably the way to go. (especially if > you want to debug on one console and run on another. sharing a keyboard > between 2 virtual consoles isn't always nice...) I suppose you have a point about the fonts, although I'm still quite happy with the fixed ones; I've used enough almost-WYSIWYG word processors! I don't think that the number of lines is too much of a restriction, provided the application is sensibly presented, and a lot of forms-based applications (typically data entry/retrieval) really tend to excel in this uncluttered environment. I'm probably biased, though, as I grew up on a Unix environment heavily hacked by IBM Mainframe veterans, so you can imagine what the user interface looked like! (I must confess to joining in with the effort, and making anything else that wasn't already forms based, even including things like getty/login, a full screen jobby) I know what you mean about the keyboard, which is why I'm using one of these MicroType things, otherwise I'd run out of space! Shame I haven't figured out what to do about the mouse yet, though. > but then a well-thought-out command line system beats the hell out of a > gui too. :> That's true. IMHO the VMS command line system is about as good as it gets, even newbies with no previous computing experience seem to pick it up very quickly. I still like forms, though! For the GUI types out there, I do quite like a well-designed graphical environment, it's just that they seem to be rather few and far between. Visual Basic is a bit of a curse, IMHO, since it allows anybody to create an insanely complex, arcane and generally unusable interface. (As I tend to give it a plug at any opportunity, the Unix based KDE project is really worth a look - one of the best things I've seen and used in ages) > and we're using dos software for internet access, having > singularly failed to set up anything else... I'd normally suggest trying one of the freebie Unices, but that one's been done to death on this newsgroup, and I'm sure if you were interested you'd have already followed it up! Chris. ###### Path: ccw.ch!usenet From: Neil.Franklin.remove.this@ccw.ch Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 22 Feb 1998 21:20:32 +0100 Organization: My own Private Self Lines: 77 Message-ID: <1zwv9ysv.fsf@chonsp.franklin.lugs.ch> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> X-Newsreader: Gnus v5.3/Emacs 19.34 On 1998-02-21 worc0223@sable.ox.ac.uk said: >:>I had a 40 MB drive on an XT, but it required partitioning because >:>32 MB was all the BIOS could handle. >:No, that's wrong. Partitioning has nothing to do with the BIOS. >:The 32 MB limitation was in DOS. The FAT16 file-system used for >:hard-disks had 16-bit block numbers, and each block stored 512 >:bytes. Therefore it could only address up to 32 MB. (The original >:FAT12 file-system used for floppies could only go up to 2 MB, >:naturally.) From: lisard@zetnet.co.uk >the 32Mb limitation was in dos, but wasn't anything to do with the fat >(which always worked in clusters; dos 1 formatted disks so that each fat >entry, or cluster, referred to 2 sectors; that allowed it to fit into a >single disk sector.) it's to do with int 0x25 and 0x26, which until dos >4 would only accept 16-bit absolute sector numbers. Also known as the DOS disk driver (is a part of IO.SYS), it simply translates to BIOS (and imposes its limits in the process of this). >the bios could >always cope with larger drives, since it used a cylinder/sector/head >numbering scheme that only finally ran out of steam when drives larger >than 4Gb came along. (how *do* you access them, please, someone?) Actually the BIOS when using WD1006 controllers (IDE drives) runs out at 504MB (1024*16*63*512, cyl*head*sect*block). With an Adaptec SCSI it will go to 1024MB (1024*64*32*512), so the top half of my 2006MB is not bootable :-( With E-IDE LBA (logical block address) controllers (2^32*512) the BIOS IMHO should pack in at 8192MB (1024*256*64*512). I actually deal with this by bypassing DOS and BIOS - with Linux, that has a 2^32 sector limit, enough for up to 2TB. That will only fail on an 5TB holostore (ca 2005..2010) :-) >fat12 could cope quite happily with drives larger than 2Mb; indeed, if >you pushed hard enough it could go right up to 32Mb - but at a cost of >using 8k clusters, which was felt to lead to too much fragmentation, so >dos 3 introduced fat16 and used 2k clusters, but only if you asked it to >deal with a partition larger than 16Mb. under that, it went happily back >to fat12, and for all we know the latest versions of dos still do. >(anybody fancy creating a 12Mb partition and testing this theory...? oh, >hang on, we'll do it ourselves.) No nead to try. I have a 10MB partition here (why waste more on DOS?): C:\>debug -l 100 2 0 1 -d 100 1421:0100 EB 34 90 4D 53 44 4F 53-33 2E 33 00 02 08 01 00 .4.MSDOS3.3..... [snipped] 1421:0170 7C 98 F7 26 16 7C 03 06-1C 7C 03 06 0E 7C A3 3F ¦..&.¦...¦...¦.? - Translate that to: EB 34 90 startup code 4D 53 44 4F 53-33 2E 33 ole MS-DOG made it 00 02 512 byte sectors 08 8 sect per cluster 01 00 1 reserved sect At 10 cylinders of 64 heads of 32 sectors = 20840 sectors this gives 20480 / 8 = 2560 clusters, just nicely in the 4096 range of FAT-12. BTW: IMHO 3.3 was the last "real" MS-DOS, 5.0 was still acceptable but not real, 6.x and 7.x were marketing (brrrr). -- private: Neil.Franklin.remove.this@ccw.ch, http://www.ccw.ch/Neil.Franklin/ for Geek Code, Papernet, Voicenet, PGP public key see http: office: franklin.remove.this@arch.ethz.ch I don't like it Microsoft, I like it Megahard If I go missing, its once again my newsfeed that has craped ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 23 Feb 1998 01:03:28 GMT Lines: 134 Message-ID: <6cqht0$qg5$2@irk.zetnet.co.uk> References: <6cpse9$4jo$2@teabag.demon.co.uk> NNTP-Posting-Host: man-094.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsfeed.wli.net!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!rill.news.pipex.net!pipex!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-22 cbh@REMOVE_THIS.teabag.demon.co.uk(ChrisHedley) said: :lisard@zetnet.co.uk writes: :> ..provided you have a high enough speed on the serial link. :>however, can you tell the terminal "you just echo everything you :>get in to the screen, and i'll read it from there, thank you"? :Well, that's why I mentioned the curses and forms packages, which :solve both the bandwidth and the "read back" problems. They keep a :copy of what's on display & the relevant attributes in memory, so hmm - what's the gain over using a memory-mapped screen then...? :> but there must be terminals which can operate in this mode, surely? doesn't the 3270 do something like this? :if you need to reference something it's dead easy, and to keep :bandwidth down they figure out how to send the absolute minimum of :data across the serial line when an update is required. Works however, we can see the utility of this - even if it doesl;ead to odd situations where the speed of screen update varies depending on how much is actually on the screen at the time. we used to program over a 9600 serial link at college, using our pc as a terminal. it was quite usable, except for that "ripple effect". :Modern dumb terminals run :at much higher speeds, in the region of 110,000bps if my memory :serves (okay, I know that 3270 terminals have run at line speeds in :the megabit range for decades now!) but that's only any use if you actually have serial ports and lines that can sustain it - using a terminal over a modem probably wouldn't run at such a speed. :> hmm. we don't mind the idea of terminals at all, except that you :>tend to be stuck with not very many lines on the screen and :>fixed-width fonts. (in our view the only good thing to come out :>of guis are proportional fonts...) but if you want to keep your :>cpu for doing cpu-y things, or want a multi-uiser box, vts are :>probably the way to go. (especially if you want to debug on one :>console and run on another. sharing a keyboard between 2 virtual :>consoles isn't always nice...) :I suppose you have a point about the fonts, although I'm still quite :happy with the fixed ones; I've used enough almost-WYSIWYG word :processors! fixed ones have their merits - speed mainly - but proportional fonts are so much easier to read that the bloat can almost be justified simply in terms of increased user pleasure. (they would make vi a pain though...) :I don't think that the number of lines is too much of a :restriction, provided the application is sensibly presented, and a :lot of forms-based applications (typically data entry/retrieval) :really tend to excel in this uncluttered environment. I'm probably that's true - forms software is pretty pleasant to use on an 80x25 format - but for programming it isn't so pleasant (word processing is somewhere between the two). for programming we like to be able to see everything that's going on at once, or as much of it as possible. this makes our coding style more horizontal than vertical, but also means that as soon as they bring out screens with the ratio of fully-extended till rolls, we're first in the queue for one. :> :biased, though, as I grew up on a Unix environment heavily hacked :by IBM Mainframe veterans, so you can imagine what the user :interface looked like! (I must confess to joining in with the :effort, and making anything else that wasn't already forms based, :even including things like getty/login, a full screen jobby) sounds like a set of ingres abf applications. :> :I know what you mean about the keyboard, which is why I'm using one :of these MicroType things, otherwise I'd run out of space! Shame I :haven't figured out what to do about the mouse yet, though. microtype things? if they're compact keyboards, please tell us more... as to the mouse - they make good cat toys, we find. :> (ours is sitting behind the mains plugs at the moment, where it has been for the last month.) :> but then a well-thought-out command line system beats the hell :>out of a gui too. :> :That's true. IMHO the VMS command line system is about as good as :it gets, even newbies with no previous computing experience seem to :pick it up very quickly. I still like forms, though! we've never used the vms command line, though there is a clone on simtel for dos that we keep meaning to investigate. how good was it? :For the GUI types out there, I do quite like a well-designed :graphical environment, it's just that they seem to be rather few :and far between. Visual Basic is a bit of a curse, IMHO, since it :allows anybody to create an insanely complex, arcane and generally :unusable interface. as are all interfaces created by people who don't have any experience or training in interface design... and microsoft's isn't terribly pretty either. there are some real minefields there. what bugs us about vb, though, is that there's no way you can script the forms. the only editor you have is the one they give you; you can't dump it to a text file, tidy it up with perl (ie. make sure everything's aligned and in the right order) and have it recompiled. grr. :(As I tend to give it a plug at any :opportunity, the Unix based KDE project is really worth a look - :one of the best things I've seen and used in ages) when we have a unix, we'll look. :> and we're using dos software for internet access, having :> singularly failed to set up anything else... :I'd normally suggest trying one of the freebie Unices, but that :one's been done to death on this newsgroup, and I'm sure if you :were interested you'd have already followed it up! oh, we're interested, we just (a) don't have a cd-rom drive, (b) have only a 386sx20 w/5Mb which we could probably use in command line mode, but it'd still be unpleasantly slow, (c) have only a 14.4k modem which is probably too slow for a net installation. so if anybody can tell us where we can get red hat on floppies... :> :> :> -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: Ingvar Mattsson Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 23 Feb 1998 12:02:48 +0100 Organization: Foreningen Lejonet, Linkoping, Sweden Lines: 19 Message-ID: References: <6cpse9$4jo$2@teabag.demon.co.uk> <6cqht0$qg5$2@irk.zetnet.co.uk> NNTP-Posting-Host: ns.idasys.se X-Newsreader: Gnus v5.3/Emacs 19.34 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!news.algonet.se!fci-se!fci!masternews.telia.net!newspost.telia.com!news.lejonet.se!not-for-mail lisard@zetnet.co.uk writes: [SNIP] > oh, we're interested, we just (a) don't have a cd-rom drive, (b) have > only a 386sx20 w/5Mb which we could probably use in command line mode, > but it'd still be unpleasantly slow, (c) have only a 14.4k modem which > is probably too slow for a net installation. > > so if anybody can tell us where we can get red hat on floppies... :> :> IMHO, any modern linux distribution is *probably* overkill on floppies, but, if you can still find it, the 1993-1994 dists fit on a rather low number on floppies (think "easily carried around"). //Ingvar (just a small, and probably useless, pointer) -- Sysadmin, disgruntled, unpolite. I don't speak for my employer nor do they speak for me. Accept this and life will be easier. ingvar@idasys.se ###### Date: 23 Feb 98 12:54:47 -0800 From: "Charlie Gibbs" Subject: Re: "640K should be enough for anyone" References: <6cpse9$4jo$2@teabag.demon.co.uk> <6cqht0$qg5$2@irk.zetnet.co.uk> Message-ID: <1260.358T2357T7746114@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 28 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!arclight.uoregon.edu!newsfeed.internetmci.com!206.229.87.25!news-peer.sprintlink.net!news-backup-west.sprintlink.net!news-in-west.sprintlink.net!news.sprintlink.net!Sprint!204.244.4.2!news.westel.com!news.skybus.com!204.244.247.121 In article <6cqht0$qg5$2@irk.zetnet.co.uk> lisard@zetnet.co.uk (lisard) writes: >On 1998-02-22 cbh@REMOVE_THIS.teabag.demon.co.uk(ChrisHedley) said: > > :lisard@zetnet.co.uk writes: > :> ..provided you have a high enough speed on the serial link. > :>however, can you tell the terminal "you just echo everything you > :>get in to the screen, and i'll read it from there, thank you"? > > :Well, that's why I mentioned the curses and forms packages, which > :solve both the bandwidth and the "read back" problems. They keep a > :copy of what's on display & the relevant attributes in memory, so > >hmm - what's the gain over using a memory-mapped screen then...? :> but >there must be terminals which can operate in this mode, surely? doesn't >the 3270 do something like this? How about using curses on a memory-mapped screen? Then you have the best of both worlds. The ancient MS-DOS version of Lattice C that I picked up had a curses library. It made a subsequent port to Unix dead easy. (The Windows port was a whole 'nother thing, though... but at least I kept keystroke equivalents so that I could preserve the feel, if not quite the look.) -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### From: cbh@REMOVE_THIS.teabag.demon.co.uk (Chris Hedley) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 23 Feb 1998 19:36:41 GMT Organization: teabag Message-ID: <6csj49$2cg$1@teabag.demon.co.uk> References: <6cpse9$4jo$2@teabag.demon.co.uk> <6cqht0$qg5$2@irk.zetnet.co.uk> NNTP-Posting-Host: localhost X-NNTP-Posting-Host: teabag.demon.co.uk [193.237.4.110] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Newsreader: knews 0.9.8 Lines: 156 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsfeed.wli.net!news1.chicago.iagnet.net!131.103.1.116!news2.chicago.iagnet.net!qual.net!iagnet.net!4.1.16.34!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!news-peer.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!news.demon.co.uk!demon!teabag.demon.co.uk!not-for-mail In article <6cqht0$qg5$2@irk.zetnet.co.uk>, lisard@zetnet.co.uk writes: > hmm - what's the gain over using a memory-mapped screen then...? :> There're a few. Firstly, because the virtual screen is free from nasty hardware related things like interrupts and all that sort of stuff (excepting the updates and so on) it puts less of a load on the OS. The relatively low bandwidth required also means a small system load, and has additional advantages such as cheap equipment, particularly for remote access, and also that a number of devices may be connected with minimal load on the system in the case of time-sharing OS'. I've used an assortment of minicomputers, generally with a fraction of the processing power of a run-of-the-mill PC of today, which had dozens of such terminals attached, which managed good response times even when they were all in use simultaneously. > but > there must be terminals which can operate in this mode, surely? Yes. One example I can think of is the Philips P7000 which used memory mapped terminals which operated across a high speed coax. These were particularly popular with word processing staff because of the phenominal (for the time) response times. The idea never really seemed to catch on, though, I'm not sure why although it may be something to do with the reasons in the above paragraph. They did do interesting things when the system was booting and didn't care much whose memory it used, though! > doesn't > the 3270 do something like this? Not as far as I'm aware (unless you're talking about the communication between the terminals and their controller, which was quite hideous - and the Unix kernel driver for ttys is apparently very similar!) They mostly just flung large records around, which appeared on the terminal as a form and to the host as a memory-accessible data entity (well, a record) > however, we can see the utility of this - even if it doesl;ead to odd > situations where the speed of screen update varies depending on how much > is actually on the screen at the time. we used to program over a 9600 > serial link at college, using our pc as a terminal. it was quite usable, > except for that "ripple effect". Many (if not most) dumb terminals have various interesting scroll commands which can be used to insert gaps or remove blocks, either on a single line or in relation to the whole screen. The terminal control packages tend to use these extensively, which cuts down on, or even entirely eliminates, this ripple-effect. > but that's only any use if you actually have serial ports and lines that > can sustain it - using a terminal over a modem probably wouldn't run at > such a speed. That's true, although a typical installation would probably involve having such terminals dotted around the same building; remote access would most likely be achieved with a network-connected terminal controller for other sites. > fixed ones have their merits - speed mainly - but proportional fonts are > so much easier to read that the bloat can almost be justified simply in > terms of increased user pleasure. (they would make vi a pain though...) I find that the fixed-font style is fairly readable, but only when applied correctly; ie. I'm using the traditional green-on-black in an x-based terminal emulator, which I find quite acceptable. On the other hand, I find that, say, black-on-white courier on the likes of Netscape is very bad for the eyes. I guess what it comes down to is what you're used to! > that's true - forms software is pretty pleasant to use on an 80x25 > format - but for programming it isn't so pleasant (word processing is > somewhere between the two). for programming we like to be able to see > everything that's going on at once, or as much of it as possible. this > makes our coding style more horizontal than vertical, but also means > that as soon as they bring out screens with the ratio of fully-extended > till rolls, we're first in the queue for one. :> There are merits to that, although I do find that a few programmers just tend to sort-of "grow into" their extra space, so they get just as little into a 50-line window that they'd previously manage on a 24 line terminal! You know the stuff, multiple blank lines (why?!), start-of-block opening brackets on a line by themselves, that sort of stuff. > sounds like a set of ingres abf applications. :> Haven't encountered that one, I've only used the CLI version of Ingres, which, as I seem to recall, was exceedingly bizarre. > microtype things? if they're compact keyboards, please tell us more... Yep, it's one of those. Don't know the exact size, I'd say about 10x6 inches. I know it sounds a bit cramped, but it's very nice to use. I understand that such things cost about 20 quid, IIRC. > as to the mouse - they make good cat toys, we find. :> (ours is sitting > behind the mains plugs at the moment, where it has been for the last > month.) Horrible, nasty things. I'm tempted to try a trackball style thing, although what sort of a state it'd be in after being used by hands which are simultaneously doing things with pizza slices is anybody's guess. > we've never used the vms command line, though there is a clone on simtel > for dos that we keep meaning to investigate. how good was it? Was? WAS?! VMS is still alive and well on the Alpha platform, and the multitude of Vaxes out there. As to how good it was, I've no doubt that the PDP-10 guys would have a lot of controvertial views, but as an OS it's brilliant. I take it you were talking about the command line, however, which really is excellent. All commands and qualifiers are logical and obvious, although for the impatient they can all be abbreviated. The CLI is consistent between the OS and its' applications, and, together with the extensive on-line help, it's a doddle to figure out what you need to type to get something done. Eg, to list files names, along with their owner and size, created before 27th January, you'd type DIRECTORY /OWNER /SIZE /BEFORE=27-JAN-1998 or, for the lazy, DIR/OW/SI/BE=27-JAN-1998 Of course the clone you mention is only likely to be for the CLI (although I guess that's enough!) and you won't get the groovy operating system with it. > as are all interfaces created by people who don't have any experience or > training in interface design... and microsoft's isn't terribly pretty > either. there are some real minefields there. This is true, although how some people manage to come up with some complete atrocities is absolutely beyond me. I know that not everyone has a flair or the training to make a good interface, but some of the stuff I've seen seems to be deliberately nasty! > what bugs us about vb, though, is that there's no way you can script the > forms. the only editor you have is the one they give you; you can't dump > it to a text file, tidy it up with perl (ie. make sure everything's > aligned and in the right order) and have it recompiled. grr. I only (briefly) used VB version 1. It was so hideous, full of bugs and inefficient that I've refused to use it since. No doubt the newer versions are better, but I'm still of the opinion that if GNU C doesn't do it, it's not worth doing! :) Of course, this newfangled Java thing has lots of interesting possibilities, provided a duff version isn't used, if you know what I'm getting at. > when we have a unix, we'll look. ... snip ... > oh, we're interested, we just (a) don't have a cd-rom drive, (b) have > only a 386sx20 w/5Mb which we could probably use in command line mode, > but it'd still be unpleasantly slow, (c) have only a 14.4k modem which > is probably too slow for a net installation. It'd run on that spec quite happily (you could even run X11, although it'd exercise your swap area, not to mention your patience!) You might be surprised at the speed, it should be a lot quicker than DOS. > so if anybody can tell us where we can get red hat on floppies... :> :> Er, there is that, I suppose. Not something I'd necessarily recommend! Chris. ###### From: Michael Kircher Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 23 Feb 1998 20:48:31 +0100 Organization: CipLab - Institutes for Physics, University of Cologne, Germany Lines: 26 Message-ID: References: <6chb3e$vod$1@nnrp1.dejanews.com> <6ci7i9$ta3$5@irk.zetnet.co.uk> NNTP-Posting-Host: jupiter.ph-cip.uni-koeln.de X-Newsreader: Gnus v5.3/Emacs 19.34 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!newsfeed.wli.net!feed2.news.erols.com!erols!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com!newsfeed.nacamar.de!news-kar1.dfn.de!news-fra1.dfn.de!news-koe1.dfn.de!RRZ.Uni-Koeln.DE!usenet lisard@zetnet.co.uk writes: > exactly what we were thinking of. think how much worse even that could > have been... ;> it's also rather amusing that the amiga managed to pull > off 32-bit sanity... btw, have the archimedes users out there been > burnt by the extension of newer arms from 26 bits to 32...? This transition didn't happen as far as user code is concerned. With the introduction of the RiscPC's Acorn used ARM6xx, ARM7xx, and StrongARM CPUs. All these CPUs feature 26bit user and privileged modes, but they issue their exceptions in 32 bit modes! All that Acorn has done is to provide wrappers that switch back into the corresponding 26-bit mode. As a consequence you aren't even *allowed* to execute 32-bit code, as the CPU could switch back due to an IRQ. Code must reside in the lower 64MB space. However data can be accessed in the entire 4GB range. This way only FIQ-handlers needed to be rewritten in 32-bit code, but this mode was seldom used in application code. So far, no one burnt :-). With Galileo, things could change. Michael ###### Path: ccw.ch!usenet From: Neil.Franklin.remove.this@ccw.ch Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 23 Feb 1998 23:03:51 +0100 Organization: My own Private Self Lines: 58 Message-ID: <67m6kmgo.fsf@chonsp.franklin.lugs.ch> References: <6cpse9$4jo$2@teabag.demon.co.uk> <6cqht0$qg5$2@irk.zetnet.co.uk> X-Newsreader: Gnus v5.3/Emacs 19.34 On 1998-02-22 cbh@REMOVE_THIS.teabag.demon.co.uk(ChrisHedley) said: :I'd normally suggest trying one of the freebie Unices, but that :one's been done to death on this newsgroup, and I'm sure if you :were interested you'd have already followed it up! lisard@zetnet.co.uk answered: >oh, we're interested, we just (a) don't have a cd-rom drive, I also only have an 3..4 year old, second hand, single speed CD-ROM. They can be had for next to nothing. Many people are pleased to get rid of them to regain the space (after upgrading to an faster one). > (b) have only a 386sx20 w/5Mb which we could probably use in command line mode, but it'd still be unpleasantly slow, Not so! I have an 386-16 (yes, pre 386DX16) with 8MB RAM. It is still surprisingly usable with Linux+XFree+xterm+Emacs+Netscape+Apache running! And that is with 1152x864 pixel with 256 colour on an non- accelerated Tseng4000-1MB video card in its ISA bus. XFree&Co was very slow in 4MB, but the console was still fast. I would expect only Linux+XFree+xterm+vi/joe to still go in 5MB. Never underestimate well written code, nor the 386 chip. You said nothing about HD space. I then had 85MB (100 - 15 (DOS)), 60 would still fit with X, 40 without. > (c) have only a 14.4k modem which is probably too slow for a net > installation. My stomach cringes at the thought. Actually if you can stand the phone bill: I have run Linux out of an 3.9MB flashdisk. As soon as my notebook is back from repair (to take a copy to work, leased line there) I would be willing to copy that partition onto my Web server for you. But I think an used CD-ROM would be cheaper that the phone bill. > so if anybody can tell us where we can get red hat on floppies... :> :> Modern distributions like Red Hat install too much unused stuff for such old machines. I would suggest you to get an ca 2 year old second hand CD and install from that or have floppies made from it. Note that the older Slackware Linux (I started with 1.0.8) distributions actually consisted of a set of floppy-sized portions on an CD. IIRC there today exists an Mini-Linux project (I have no URL, search the Web). That would be floppyable, perhaps even managable on a modem. -- private: Neil.Franklin.remove.this@ccw.ch, http://www.ccw.ch/Neil.Franklin/ for Geek Code, Papernet, Voicenet, PGP public key see http: office: franklin.remove.this@arch.ethz.ch I don't like it Microsoft, I like it Megahard If I go missing, its once again my newsfeed that has craped ###### From: amlaukka@cc.helsinki.fi (Aki M Laukkanen) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 24 Feb 1998 08:41:21 GMT Organization: University of Helsinki Lines: 11 Message-ID: References: <6cfi0b$q9o$4@irk.zetnet.co.uk> <6chb3e$vod$1@nnrp1.dejanews.com> NNTP-Posting-Host: vesuri.helsinki.fi X-Newsreader: slrn (0.9.3.2 UNIX) Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!feeder.qis.net!nntp.abs.net!newsfeed1.telenordia.se!newsfeed1.funet.fi!news.funet.fi!news.helsinki.fi!amlaukka In article <6chb3e$vod$1@nnrp1.dejanews.com>, jk@langley.softwright.co.uk wrote: >Does anyone remember if any of the major apps got hit on 32-bit >compatibility? None spring to mind right now, but I've been away from the Mac >world for several years... Well, this is not a Macintosh application but the AmigaBasic by Microsoft didn't do well with 32-bit addressbus processors. -- D. ###### From: Robert Billing Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Tue, 24 Feb 98 09:14:20 GMT Message-ID: <888311660snz@tnglwood.demon.co.uk> References: <6cpse9$4jo$2@teabag.demon.co.uk> <6csj49$2cg$1@teabag.demon.co.uk> Reply-To: unclebob@tnglwood.demon.co.uk X-Mail2News-User: unclebob@tnglwood.demon.co.uk X-Mail2News-Path: post-10.mail.demon.net!post.mail.demon.net!tnglwood.demon.co.uk X-Trace: mail2news.demon.co.uk 888345388 14114 unclebob tnglwood.demon.co.uk X-Complaints-To: abuse@demon.net X-Newsreader: Demon Internet Simple News v1.29 Lines: 26 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!nntp.flash.net!207.114.4.11.MISMATCH!nntp.abs.net!news-dc.gip.net!news-peer.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!tnglwood.demon.co.uk!unclebob In article <6csj49$2cg$1@teabag.demon.co.uk> cbh@REMOVE_THIS.teabag.demon.co.uk "Chris Hedley" writes: > type to get something done. Eg, to list files names, along with their > owner and size, created before 27th January, you'd type > DIRECTORY /OWNER /SIZE /BEFORE=27-JAN-1998 > or, for the lazy, DIR/OW/SI/BE=27-JAN-1998 DCL is in fact so intuitive that it is possible to make jokes in it. A few years ago the UK DECUS newsletter ran a caption competition, the picture being of a row of VAX cabinets with a big puddle in front of them. The winning entry was SET DEVICE /DEFROST /OUTPUT=FLOOR but apart from that, having migrated to a world in which I now use a flock of unices, I miss the regularity that comes from the way in which the DCL shell and the applications interacted. -- I am Robert Billing, Christian, inventor, traveller, cook and animal lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/ "Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock phasers on the Heffalump, Piglet, meet me in transporter room three" ###### From: benc@dcs.qmw.ac.uk (Ben Clifford) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 24 Feb 1998 09:33:41 GMT Organization: Queen Mary & Westfield College, London, UK Lines: 24 Message-ID: <6cu45l$rnr$1@beta.qmw.ac.uk> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> NNTP-Posting-Host: jason.dcs.qmw.ac.uk X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!server6.netnews.ja.net!server4.netnews.ja.net!qmw!benc lisard@zetnet.co.uk wrote: : the bios could : always cope with larger drives, since it used a cylinder/sector/head : numbering scheme that only finally ran out of steam when drives larger : than 4Gb came along. (how *do* you access them, please, someone?) Don't use the BIOS. The BIOS disk routines are just a few driver routines for accessing the drive hardware - you can write your own routines that do the same and they can do whatever you want. I presume that most OSs do that anyway as I expect the BIOS routines only work in real mode. OS/2 certainly had the option to use the BIOS for disk access so I guess that when you weren't using this option, it hit the hardware directly. -- I didn't believe in God, but then I discovered I am Him. http://www.dcs.qmw.ac.uk/~benc/ ###### From: die.spam@hell.org.su (Evandro Menezes) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Tue, 24 Feb 1998 21:35:55 GMT Organization: E&R Lines: 17 Message-ID: <34f435ad.4507066@news.nabi.net> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> Reply-To: evandro@geocities.com (Evandro Menezes) NNTP-Posting-Host: 24818@208.6.184.182 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!news.ecn.uoknor.edu!news.wildstar.net!newsfeed.direct.ca!Supernews60!supernews.com!Supernews69!not-for-mail On 22 Feb 1998 00:28:54 GMT, lisard@zetnet.co.uk wrote: >the *real* problem with the fat is that it's basically a linear access >system; to get to the last cluster in your file, you have to read all >the fat entries in. great, if they're all together; tough, if they're >scattered across the length of your hard drive... The FAT sectors are ALWAYS laid down consecutively on the disk. BTW, DOS has always kept the FAT "cached". Though the FAT on the disk had to be accessed in order to be updated. HTH _______________________________________________________________ Evandro Menezes Austin, TX USA Tel:+1-512-502-9199 ICQ:7957253 evandro@geocities.com www.geocities.com/SiliconValley/7263 ###### From: kmurcray@$TIN_DOMAIN (PEREGRINE) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 24 Feb 1998 22:21:43 GMT Organization: University of Denver / Physics Dept Lines: 15 Message-ID: <6cvh5n$s8k@undetermined> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> <6cu45l$rnr$1@beta.qmw.ac.uk> Reply-To: kmurcray@du.edu NNTP-Posting-Host: diac.com X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!news.ecn.uoknor.edu!feed1.news.erols.com!howland.erols.net!cpk-news-hub1.bbnplanet.com!denver-news-feed1.bbnplanet.com!news.bbnplanet.com!news.diac.com!kmurcray Ben Clifford (benc@dcs.qmw.ac.uk) wrote: : OS/2 certainly had the option to use the BIOS for disk access so I : guess that when you weren't using this option, it hit the hardware : directly. Had? HAD? Why does everybody keep talking about OS/2 in past tense? >80% of our machines here at work are running Warp version 4.0... -- Kevin G. Murcray kmurcray@du.edu University of Denver Physics & Astronomy http://www.du.edu/~kmurcray/ Atmospheric Research Group True Ghost Stories: http://www.du.edu/~kmurcray/ghost.html ###### From: benc@dcs.qmw.ac.uk (Ben Clifford) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 24 Feb 1998 22:56:01 GMT Organization: Queen Mary & Westfield College, London, UK Lines: 12 Message-ID: <6cvj61$g30$1@beta.qmw.ac.uk> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> <6cu45l$rnr$1@beta.qmw.ac.uk> <6cvh5n$s8k@undetermined> NNTP-Posting-Host: jason.dcs.qmw.ac.uk X-Newsreader: TIN [version 1.2 PL2] Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!server6.netnews.ja.net!server4.netnews.ja.net!qmw!benc PEREGRINE (kmurcray@$TIN_DOMAIN) wrote: : Had? HAD? Why does everybody keep talking about OS/2 in past tense? >80% : of our machines here at work are running Warp version 4.0... I use "had" because I don't run it any more: I moved over to linux because I wanted a change. -- I didn't believe in God, but then I discovered I am Him. http://www.dcs.qmw.ac.uk/~benc/ ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 24 Feb 1998 23:53:24 GMT Message-ID: <6cvmhk$1de$6@irk.zetnet.co.uk> References: <6csj49$2cg$1@teabag.demon.co.uk> NNTP-Posting-Host: man-164.dialup.zetnet.co.uk X-Everything: Net-Tamer V 1.08X Lines: 165 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!newsfeed.usit.net!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!nntp.news.xara.net!xara.net!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-02-23 cbh@REMOVE_THIS.teabag.demon.co.uk said: :The relatively low bandwidth required also means a small :system load, and has additional advantages such as cheap equipment, :particularly for remote access, and also that a number of devices :may be connected with minimal load on the system in the case of :time-sharing OS'. but surely you're still stuck with the basic requirement to set aside 2k per terminal to take care of the virtual screen? if you've got 16 terminals on a 64k system, that's half your memory gone. anyway we were thinking more of a terminal where you could just send a request to it saying "tell me what's on the screen at position (x,y)" - or "between positions (x,y) and (z,t) and you worry about the keyboard from now on" :> however, we can see the utility of this - even if it does lead to :>odd situations where the speed of screen update varies depending :>on how much is actually on the screen at the time. we used to :>program over a 9600 serial link at college, using our pc as a :>terminal. it was quite usable, except for that "ripple effect". :Many (if not most) dumb terminals have various interesting scroll :commands which can be used to insert gaps or remove blocks, either :on a single line or in relation to the whole screen. The terminal :control packages tend to use these extensively, which cuts down on, :or even entirely eliminates, this ripple-effect. maybe we just had dodgy vt100 termcaps, but we didn't see much of this. on the other hand, we did tend to scroll through source files a lot, and full-page refreshes over a 9600 link are never going to take less than a second on the average - which is quite high when you're used to pc editors that refresh in a couple of frames, even when you're using an xt. :I find that the fixed-font style is fairly readable, but only when :applied correctly; ie. I'm using the traditional green-on-black in :an x-based terminal emulator, which I find quite acceptable. On :the other hand, I find that, say, black-on-white courier on the :likes of Netscape is very bad for the eyes. I guess what it comes :down to is what you're used to! probably. we're used to fixed fonts, though, having spent most of our life with computers using hardware that couldn't or didn't support proportional ones. what would be nice would be a terminal with support for proportional fonts on an otherwise terminal-ish setup. (well, just the one. just so that width(i) < width(x) < width(m). ) :There are merits to that, although I do find that a few programmers :just tend to sort-of "grow into" their extra space, so they get :just as little into a 50-line window that they'd previously manage :on a 24 line terminal! You know the stuff, multiple blank lines :(why?!), start-of-block opening brackets on a line by themselves, :that sort of stuff. we know what you mean here. we always bunch our braces, and closing braces tend to go on the same line as the last line of code. what would be nice is a control code that emitted half a blank line, since often for separating procedures, a whole one is too much. :> microtype things? if they're compact keyboards, please tell us :>more... :Yep, it's one of those. Don't know the exact size, I'd say about :10x6 inches. I know it sounds a bit cramped, but it's very nice to :use. I understand that such things cost about 20 quid, IIRC. where can we get one, and do they work with pcs? it sounds a bit cramped, but we grew up on a zx81, so we'll survive. :> :> as to the mouse - they make good cat toys, we find. :> (ours is :>sitting behind the mains plugs at the moment, where it has been :>for the last month.) :Horrible, nasty things. hear hear! (though not as nasty as applications that assume you (a) have one, (b) prefer it to the keyboard, and (c) don't mind dropping out of work in a few years' time because your right hand is RSI'd beyond further use...) :I'm tempted to try a trackball style thing, :although what sort of a state it'd be in after being used by hands :which are simultaneously doing things with pizza slices is :anybody's guess. we'd like to find a driver that will let us use the numeric keypad as a mouse, so that there are (essentially) 2 key-driven cursors on the screen at once. :> we've never used the vms command line, though there is a clone on :>simtel for dos that we keep meaning to investigate. how good was :>it? :Was? WAS?! excuse us whilst we wipe our feet. ;> is. sorry. :VMS is still alive and well on the Alpha platform, and :the multitude of Vaxes out there. ah, but we use alphas at work - look systems is a digital var - and we've never seen vms... is it openvms now? and how open is it? (ie. source code?) and is bliss still used as a programming language...? :As to how good it was, I've no :doubt that the PDP-10 guys would have a lot of controvertial views, :but as an OS it's brilliant. I take it you were talking about the :command line, however, which really is excellent. *nod* :All commands and :qualifiers are logical and obvious, although for the impatient they :can all be abbreviated. The CLI is consistent between the OS and :its' applications, and, together with the extensive on-line help, :it's a doddle to figure out what you need to type to get something :done. lovely. its applications...? please describe. :Of course the clone you mention is only likely to be for the CLI :(although I guess that's enough!) and you won't get the groovy :operating system with it. shame, that, but one can't have everything in this life. :I only (briefly) used VB version 1. It was so hideous, full of :bugs and inefficient that I've refused to use it since. No doubt :the newer versions are better, but I'm still of the opinion that if :GNU C doesn't do it, it's not worth doing! :) well, we've only used vb5cce, so we can't really say much, but we wouldn't like to have to program applications with a lot of forms or dialog boxes with it. :Of course, this :newfangled Java thing has lots of interesting possibilities, :provided a duff version isn't used, if you know what I'm getting at. it may have lots of interesting possibilities, but it won't catch on until it reconciles itself better with the windows way of doing things. the big problem with java is the gui performance; raw speed is quite acceptable. of course, this all relates to sun's win95 implementation of java 1.1 and 1.0.2 - microsoft's implementations are trivial and shite. :> oh, we're interested, we just (a) don't have a cd-rom drive, (b) :>have only a 386sx20 w/5Mb which we could probably use in command :>line mode, but it'd still be unpleasantly slow, (c) have only a :>14.4k modem which is probably too slow for a net installation. :It'd run on that spec quite happily (you could even run X11, :although it'd exercise your swap area, not to mention your :patience!) You might be surprised at the speed, it should be a lot :quicker than DOS. ah. we'll have to go out and buy a cd-rom then... -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... Net-Tamer V 1.08X - Test Drive ###### From: lduning@mdisystems.com (Lars Duning) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: Thu, 26 Feb 1998 10:23:58 GMT Organization: BT Internet Lines: 26 Message-ID: <34f541e6.3065600@news.btinternet.com> References: <6cfi0b$q9o$4@irk.zetnet.co.uk> <6chb3e$vod$1@nnrp1.dejanews.com> Reply-To: lars@cableinet.co.uk NNTP-Posting-Host: host5-99-61-182.btinternet.com X-Newsreader: Forte Free Agent 1.11/32.235 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!feeder.qis.net!btnet-peer!btnet!neptunium.btinternet.com!not-for-mail On Wed, 25 Feb 1998 11:22:52 +0100, peterk@combo.ganesha.com (Dr. Peter Kittel) wrote: >In article amlaukka@cc.helsinki.fi (Aki M Laukkanen) writes: >>In article <6chb3e$vod$1@nnrp1.dejanews.com>, jk@langley.softwright.co.uk wrote: >> >>>Does anyone remember if any of the major apps got hit on 32-bit >>>compatibility? None spring to mind right now, but I've been away from the Mac >>>world for several years... >> >>Well, this is not a Macintosh application but the AmigaBasic by Microsoft >>didn't do well with 32-bit addressbus processors. > >Indeed, but in fact *because* it was a Macintosh application in its >former life :-) ! As I heard it, AmigaBasic was a direct port of >Microsoft's Mac Basic, relying on a 24-bit address space... >So here it said "16 MB should be enough for anyone". Maybe MS thought that Motorola would take the hint and, like Intel[1], change the processor specs according to MS' programming, not vice versa. Fortunately, Motorola didn't. [1] Have a nice article about it.... somewhere in my boxes... I hate moving! -- Lars Duening; lars@cableinet.co.uk (Home) ###### From: ascroff@mdiu.edu (Andrew Scroffit) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 26 Feb 1998 15:34:53 GMT Organization: Maryland Inland University Lines: 30 Message-ID: <888507393.87840@ins8.netins.net> References: <6cfi0b$q9o$4@irk.zetnet.co.uk> <6chb3e$vod$1@nnrp1.dejanews.com> NNTP-Posting-Host: ins8.netins.net Mime-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII X-Newsreader: WinVN 0.99.8 (x86 32bit) Cache-Post-Path: ins8.netins.net!unknown@desm-28-38.dialup.netins.net Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!news-feed.fnsi.net!news.idt.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.internetmci.com!167.142.225.6!newsrelay.netins.net!news.netins.net!not-for-mail In article , peterk@combo.ganesha.com says... > >In article amlaukka@cc.helsinki.fi > (Aki M Laukkanen) writes: >>In article <6chb3e$vod$1@nnrp1.dejanews.com>, jk@langley.softwright.co.uk wrote >: >> >>>Does anyone remember if any of the major apps got hit on 32-bit >>>compatibility? None spring to mind right now, but I've been away from the Mac >>>world for several years... >> >>Well, this is not a Macintosh application but the AmigaBasic by Microsoft >>didn't do well with 32-bit addressbus processors. > >Indeed, but in fact *because* it was a Macintosh application in its >former life :-) ! As I heard it, AmigaBasic was a direct port of >Microsoft's Mac Basic, relying on a 24-bit address space... >So here it said "16 MB should be enough for anyone". > Weren't the first Macs with Nubus memory mapped into 4 4M regions? 4M reserved for NUBUS 4M reserved for ROM 4M for user 4M for ? (I forget now, the OS? or screen?) Andy ###### Path: ccw.ch!usenet From: Neil.Franklin.remove.this@ccw.ch Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 26 Feb 1998 22:11:22 +0100 Organization: My own Private Self Lines: 21 Message-ID: References: <6cfi0b$q9o$4@irk.zetnet.co.uk> <6chb3e$vod$1@nnrp1.dejanews.com> <888507393.87840@ins8.netins.net> X-Newsreader: Gnus v5.3/Emacs 19.34 ascroff@mdiu.edu (Andrew Scroffit) asked: >Weren't the first Macs with Nubus memory mapped into 4 4M regions? >4M reserved for NUBUS >4M reserved for ROM >4M for user >4M for ? (I forget now, the OS? or screen?) AFAIK (not much) it was: 0..7 8M RAM 8 1M Motherboard IO 9..E 6*1M Nubus Slots (thats why there was max 6 (MAC II, IIx, IIfx)) F 1M ROM -- private: Neil.Franklin.remove.this@ccw.ch, http://www.ccw.ch/Neil.Franklin/ for Geek Code, Papernet, Voicenet, PGP public key see http: office: franklin.remove.this@arch.ethz.ch I don't like it Microsoft, I like it Megahard If I go missing, its once again my newsfeed that has craped ###### From: cbh@REMOVE_THIS.teabag.demon.co.uk (Chris Hedley) Newsgroups: alt.folklore.computers Subject: Re: "640K should be enough for anyone" Date: 26 Feb 1998 22:12:32 GMT Organization: teabag Message-ID: <6d4pcg$58m$1@teabag.demon.co.uk> References: <6cnj4e$3db$1@eldritch.dyn.ml.org> <6cnrg6$3us$1@irk.zetnet.co.uk> <1zwv9ysv.fsf@chonsp.franklin.lugs.ch> NNTP-Posting-Host: localhost X-NNTP-Posting-Host: teabag.demon.co.uk [193.237.4.110] Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Newsreader: knews 0.9.8 Lines: 12 Path: ccw.ch!aetna.dolphins.ch!news.planetc.com!leto.ou.edu!hammer.uoregon.edu!logbridge.uoregon.edu!news.maxwell.syr.edu!woodstock.news.demon.net!demon!news.demon.co.uk!demon!teabag.demon.co.uk!not-for-mail In article <1zwv9ysv.fsf@chonsp.franklin.lugs.ch>, Neil.Franklin.remove.this@ccw.ch writes: > I actually deal with this by bypassing DOS and BIOS - with Linux, that > has a 2^32 sector limit, enough for up to 2TB. That will only fail on an > 5TB holostore (ca 2005..2010) :-) By which time the older versions of Linux will have a 256 bit address bus, so I think your holostore should be okay on that front. Unfortunately, all those high sample rate ultra-resolution 3d video compatible files will still fill up your storage. Such is progress, I suppose. Chris.