Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Fortran source as instruction code (was Re: Rubout) Date: 11 Oct 1998 23:04:28 +0200 Organization: My own Private Self Lines: 22 Sender: neil@chonsp.franklin.ch Message-ID: References: <908089676@ablelink.org> X-Newsreader: Gnus v5.3/Emacs 19.34 justin.frim@ablelink.org (Justin Frim) writes: > > SEEBS@PLETHORA.NET wrote: > > >Anything that is intended > >to be manipulated by people should be readable. > > Then we would still be running source code ROM BASIC programs instead > of machine code, because according to your argument, if people have to This reminded me of an project (IIRC in the 70s) to build an computer that directly ran Fortran (II?, 66?, 77?) source as its instruction code. Does anyone know any details about this? Was it ever built or was it just an off beat idea? -- *** New home Addresses Mail and Web *** home: neil@franklin.ch.remove http://neil.franklin.ch/ work: franklin@arch.ethz.ch.remove http://caad.arch.ethz.ch/~franklin/ Microsoft is Software Communism, Fight for GNU Freedom! ###### From: jones@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 11 Oct 1998 22:06:20 GMT Organization: The University of Iowa Lines: 16 Message-ID: <6vra4s$no4$1@flood.weeg.uiowa.edu> References: NNTP-Posting-Host: pyrite.cs.uiowa.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!news.maxwell.syr.edu!newsfeed.cwix.com!128.174.5.49!vixen.cso.uiuc.edu!NewsNG.Chicago.Qual.Net!news.uiowa.edu!not-for-mail From article , by Neil Franklin : > > This reminded me of an project (IIRC in the 70s) to build an computer > that directly ran Fortran (II?, 66?, 77?) source as its instruction code. The Fairchild Symbol Machine, built at Iowa State University, directly executed the Symbol programming language. This was a marvelous dead-end in computer architecture. All functions we now think of as system software functions were done in hardware: The OS, the source-code text editor, and the compiler were dedicated hardware. If I remember correctly, the text editor did somewhat of a byte-code reduction of the source program, and the source text was reconstructed from this when you asked for a listing. You could, therfore, refer to Symbol as relying on incremental compilation. Doug Jones jones@cs.uiowa.edu ###### From: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 11 Oct 1998 23:19:14 GMT Organization: The National Capital FreeNet Lines: 11 Message-ID: <6vredi$roc@freenet-news.carleton.ca> References: <908089676@ablelink.org> Reply-To: ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) NNTP-Posting-Host: freenet3.carleton.ca X-Given-Sender: ab528@freenet3.carleton.ca (Heinz W. Wiggeshoff) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!europa.clark.net!209.130.129.214!nntp.frontiernet.net!node17.frontiernet.net!xcski.com!freenet-news.carleton.ca!FreeNet.Carleton.CA!ab528 Neil Franklin (neil@franklin.ch.remove) writes: > > This reminded me of an project (IIRC in the 70s) to build an computer > that directly ran Fortran (II?, 66?, 77?) source as its instruction code. > > Does anyone know any details about this? Was it ever built or was it > just an off beat idea? APL and LISP were more suitable candidates. And yes, and (marketing- wise) yes for A&L. ###### From: mlevine@ridgecrest.ca.us (Michael N. LeVine) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Sun, 11 Oct 1998 19:21:53 -0700 Organization: RidgeNET Lines: 25 Message-ID: References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> NNTP-Posting-Host: pm165.ridgenet.net X-Newsreader: MT-NewsWatcher 2.4.4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!199.120.150.5!news.ridgecrest.ca.us!mlevine In article <6vredi$roc@freenet-news.carleton.ca>, ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) wrote: > Neil Franklin (neil@franklin.ch.remove) writes: > > > > This reminded me of an project (IIRC in the 70s) to build an computer > > that directly ran Fortran (II?, 66?, 77?) source as its instruction code. > > > > Does anyone know any details about this? Was it ever built or was it > > just an off beat idea? > > APL and LISP were more suitable candidates. And yes, and (marketing- > wise) yes for A&L. IIRC this sounds a lot like the effort in the 70's to build a machine that would run P-code (from Wirths Pascal Compiler) directly. Remember UCSD Pascal operating system ?? It was a software virtual machine that could be ported to any hardware platform and run the same programs. A similar concept as that of the Java virtual machine. -- Michael LeVine - mlevine@ridgecrest.ca.us "Thirty days hath September, April, June and November. All the rest have thirty one except for Gypsy Rose Lee and every one knew what she had" - Mel Blanc ###### From: dpeschel@u.washington.edu (D. Peschel) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 12 Oct 1998 03:36:49 GMT Organization: University of Washington, Seattle Lines: 23 Message-ID: <6vrtgh$16es$1@nntp1.u.washington.edu> References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> NNTP-Posting-Host: saul4.u.washington.edu X-Trace: nntp1.u.washington.edu 908163409 39388 (None) 140.142.17.38 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: dpeschel Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsfeed.cwix.com!192.220.250.21!netnews1.nw.verio.net!netnews.nwnet.net!news.u.washington.edu!dpeschel In article , Michael N. LeVine wrote: >In article <6vredi$roc@freenet-news.carleton.ca>, >ab528@FreeNet.Carleton.CA (Heinz W. Wiggeshoff) wrote: >> Neil Franklin (neil@franklin.ch.remove) writes: >> > This reminded me of an project (IIRC in the 70s) to build an computer >> > that directly ran Fortran (II?, 66?, 77?) source as its instruction code. >IIRC this sounds a lot like the effort in the 70's to build a machine that >would run P-code (from Wirths Pascal Compiler) directly. Remember UCSD >Pascal operating system ?? It was a software virtual machine that could be >ported to any hardware platform and run the same programs. A similar >concept as >that of the Java virtual machine. It may sound like it but they're not at all the same thing. You're talking about a machine that directly supports the compiled version of a language. (The fact that this machine is virtual is irrelevant.) The original poster was talking about a machine that directly supports the UNCOMPILED version of a language. -- Derek ###### From: "George R. Gonzalez" Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Mon, 12 Oct 1998 13:46:32 -0500 Organization: University of Minnesota, Twin Cities Campus Lines: 26 Message-ID: <6vtj0o$8t8$1@news1.tc.umn.edu> References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> NNTP-Posting-Host: grg1.micro.umn.edu Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Newsreader: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!enews.sgi.com!news.eecs.umich.edu!newshub.tc.umn.edu!news1.tc.umn.edu!not-for-mail There's several difficulties in answering this question. First of all, it's not generally possible to just start executing source code. There are problems with forward references-- just think of all the GOTO's in your typical FORTRAN program. So it really can't be a pure interpreter-- it's gotta have some pre-passes to figure out what jumps where. ... And if it's doing that, it's tempting to do a bit more in that pass, such as build tokens, parse expressions,... and then you've done about half of a compiler's job. Next, it's unlikely anybody could ever design real hardware to do the job. It would take a heck of a lot of NAND gates! What has been done, quite a few times, is to write an interpreter in microcode which appears to act like hardware. But of course it's just software, hidden a bit. Thirdly, it's not really cost effective to put a HUGE interpreter in hardware. The current schemes of pre-compiling to some intermediate code or all the way to machine code are much more economical ways to go. You may find references to many PhD theses with titles like "APL in hardware", but these are almost certainly blue-sky exercises-- not likely anything anybody has actually implemented or marketed. ###### From: rivie@rivie.daautah.com (Roger Ivie) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 12 Oct 1998 20:22:57 GMT Organization: XMission http://www.xmission.com/ Lines: 22 Message-ID: References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> Reply-To: rivie@daa-utah.com NNTP-Posting-Host: logan30.modem.xmission.com X-Trace: news.xmission.com 908223777 21899 166.70.3.222 (12 Oct 1998 20:22:57 GMT) X-Complaints-To: abuse@xmission.com NNTP-Posting-Date: 12 Oct 1998 20:22:57 GMT X-Newsreader: slrn (0.8.8.2 UNIX) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!xmission!nnrp.xmission!rivie In article <6vtj0o$8t8$1@news1.tc.umn.edu>, George R. Gonzalez wrote: >There's several difficulties in answering this question. I missed the question. Sorry. > >Next, it's unlikely anybody could ever design real hardware to do the >job. It would take a heck of a lot of NAND gates! What has been done, >quite a few times, is to write an interpreter in microcode which appears >to act like hardware. But of course it's just software, hidden a bit. The famous Bell&Newell computer architecture book describes a machine designed to execute FORTRAN II in hardware. IIRC, it did a small bit of tokenization as the source code was read in, but not as much as (say) the old Microsoft BASICs. I don't have my copy handy. -- Roger Ivie Design Analysis Associates 75 West 100 South Logan, UT 84321 mailto:rivie@daa-utah.com phoneto:(435)753-2212 ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Rubout) Date: 13 Oct 1998 20:32:14 GMT Lines: 15 Message-ID: <700dce$950$1@irk.zetnet.co.uk> References: <6vtj0o$8t8$1@news1.tc.umn.edu> NNTP-Posting-Host: man-138.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 908310734 9376 194.247.40.176 (13 Oct 1998 20:32:14 GMT) NNTP-Posting-Date: 13 Oct 1998 20:32:14 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!newscore.univie.ac.at!btnet-peer!btnet!news.freedom2surf.net!ayres.ftech.net!news.ftech.net!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-10-12 grg@foundsys.com said: :You may find references to many PhD theses with titles like "APL in :hardware", but these are almost certainly blue-sky exercises-- not :likely anything anybody has actually implemented or marketed. Of course, there have been a fair few attempts at putting Forth in hardware, with various degrees of commercial success. Not generally including the text interpreter, but Forth's intermediate representation (pure threaded code) does lend itself well to direct-to-hardware representation. -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... ###### Path: chonsp.franklin.ch!usenet From: Neil Franklin Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 13 Oct 1998 22:52:01 +0200 Organization: My own Private Self Lines: 69 Sender: neil@chonsp.franklin.ch Message-ID: References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> X-Newsreader: Gnus v5.3/Emacs 19.34 "George R. Gonzalez" writes: > > First of all, it's not generally possible to just start executing source > code. Actually one can immediately interpret source with an interpreter written in software. BASIC only used tokens to save memory and execution time on small microprocessors and for higher performance today, not to make running possible. So strictly microcode should also be able to do it. Of course that does not mean that it will be fast. > There are problems with forward references-- just think of all the GOTO's > in your typical FORTRAN program. After hitting an GOTO, do an forward scan in the source. Slow but possible. And no doubt you can build an list of pointers to lines or even to labels. Quasy dynamic optimising. Again: it may be slow and eat pointer cache space for lunch, but it is possible. > So it really can't be a pure > interpreter-- it's gotta have some pre-passes to figure out what jumps > where. If you regard passes as parallel processes and timeslice them and then ditch the generic timeslicer for explicit jumping between different states, then you have got an pure interpreter again. > Next, it's unlikely anybody could ever design real hardware to do the > job. It would take a heck of a lot of NAND gates! What has been done, > quite a few times, is to write an interpreter in microcode which appears > to act like hardware. But of course it's just software, hidden a bit. I was assuming an microcoded system. After all a 80x86 interprets its instruction code only partially in pure hardware, the rest in microcode. > Thirdly, it's not really cost effective to put a HUGE interpreter in > hardware. The current schemes of pre-compiling to some intermediate code > or all the way to machine code are much more economical ways to go. I was not asking about efficiency (I am an RISC fan). I was more trying to determine whether the "Fortran in hardware" stories I have heard were urban legends of the "CISC at the most ridiculous" variety. Or whether someone had actually designed or even implemented such a thing. > You may find references to many PhD theses with titles like "APL in > hardware", but these are almost certainly blue-sky exercises-- not likely > anything anybody has actually implemented or marketed. Those I know of these were allways Lisp/APL/Forth optimised token interpreters. Not source interpreters. And some of them (at least for Lisp and Forth) have been successfully produced and sold. -- *** New home Addresses Mail and Web *** home: neil@franklin.ch.remove http://neil.franklin.ch/ work: franklin@arch.ethz.ch.remove http://caad.arch.ethz.ch/~franklin/ Microsoft is Software Communism, Fight for GNU Freedom! ###### Message-ID: <3624C81A.658E@gazonk.del> Date: Wed, 14 Oct 1998 11:49:46 -0400 From: "Foobar T. Clown" Reply-To: foobar@gazonk.del Organization: Blurp X-Mailer: Mozilla 3.01 (WinNT; I) MIME-Version: 1.0 Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 198.133.110.49 X-Trace: 14 Oct 1998 11:49:54 -0500, 198.133.110.49 Lines: 11 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!newspeer.monmouth.com!ix.netcom.com!zeus.nomos.com!198.133.110.49 Neil Franklin wrote: > > Actually one can immediately interpret source with an interpreter > written in software. [...] So strictly microcode should also be able > to do it. Microcode *IS* software. Don't be fooled by the fact that it is often written by people who call themselves "Hardware Engineers," or by the fact that they describe its function as being "done by the hardware." That's just because they don't want any software geeks stepping on their turf. ###### From: "George R. Gonzalez" Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Wed, 14 Oct 1998 11:17:48 -0500 Organization: SkyPoint Communications, Inc. Lines: 33 Message-ID: <702j2m$6um$1@shadow.skypoint.net> References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> <3624C81A.658E@gazonk.del> NNTP-Posting-Host: dial051.skypoint.net Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Newsreader: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!newsfeed.cwix.com!199.170.121.3!falcon.america.net!news.pagesat.net!skypoint.com!not-for-mail Foobar T. Clown wrote in message <3624C81A.658E@gazonk.del>... >Neil Franklin wrote: >> >> Actually one can immediately interpret source with an interpreter >> written in software. [...] So strictly microcode should also be able >> to do it. > >Microcode *IS* software. Don't be fooled by the fact that it is often >written by people who call themselves "Hardware Engineers," or by the >fact that they describe its function as being "done by the hardware." >That's just because they don't want any software geeks stepping on their >turf. Good point. The line between hardware, microcode, and machine code is a very fuzzy one. It gets even fuzzier if you realize that the lowest level of "microcode" is begin interpreted by hardware that nowdays is often described in some logic design language. The logic design language gets compiled into a set of logic maps that get burned into a gate array chip. So in some sense even the doggone microcode is being interpreted by a program! I'll have to stick to an old conservative position-- It's only a true FORTRAN computer if there are NAND/NOR gates and flip-flops wired together to do the whole job, and not a single line of "programming", no ROMs or FPLA's. I'll allow multiplexers and RAM though.... It should be no bigger than a SAGE computer, and probably a good use for all those unemployed Soviet engineers.... ###### From: glass2@glass2.cv.lexington.ibm.com Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 14 Oct 1998 16:25:01 GMT Organization: IBM Austin Lines: 32 Message-ID: <702j8t$10m4$1@ausnews.austin.ibm.com> References: <908089676@ablelink.org> <3624C81A.658E@gazonk.del> Reply-To: wa4qal@vnet.ibm.com NNTP-Posting-Host: glass2.cv.lexington.ibm.com X-Newsreader: IBM NewsReader/2 2.0 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!Supernews60!supernews.com!uunet!in4.uu.net!an02.austin.ibm.com!ausnews.austin.ibm.com!not-for-mail In <3624C81A.658E@gazonk.del>, "Foobar T. Clown" writes: >Neil Franklin wrote: >> >> Actually one can immediately interpret source with an interpreter >> written in software. [...] So strictly microcode should also be able >> to do it. > >Microcode *IS* software. Don't be fooled by the fact that it is often >written by people who call themselves "Hardware Engineers," or by the >fact that they describe its function as being "done by the hardware." >That's just because they don't want any software geeks stepping on their >turf. I seem to remember one machine, was it the 5500, that had custom microcode for different languages. Thus, to run a Fortran application, it would load the Fortran microcode. To run a Cobol application, it would load the Cobol microcode. This allowed the microcode to be customized to the particular application. I don't remember if the machine was designed to be a multitasking machine or not. Still, it's possible that the microcode could be pagable at task swap time. Pagable microcode was also used in the IBM 308x series of machines. A section of main storage was partitioned off, and was used to hold part of the microcode. I'm still waiting for user programmable microcode on a microcomputer (Ok, so I'm a "Hardware Engineer" as well as a "Programmer".). :*) Dave P.S. Standard Disclaimer: I work for them, but I don't speak for them. ###### From: hnsngr@sirius.com (Ron Hunsinger) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Wed, 14 Oct 1998 17:07:36 -0700 Organization: ErsteSoft Lines: 13 Message-ID: References: <908089676@ablelink.org> <3624C81A.658E@gazonk.del> <702j8t$10m4$1@ausnews.austin.ibm.com> NNTP-Posting-Host: ppp-asok02--062.sirius.net Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: Yet Another NewsWatcher 2.3.1 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.shore.net!uunet!in4.uu.net!newshub.sirius.com!newsfiler.sirius.com!hnsngr In article <702j8t$10m4$1@ausnews.austin.ibm.com>, wa4qal@vnet.ibm.com wrote: > I seem to remember one machine, was it the 5500, that had custom microcode > for different languages. Thus, to run a Fortran application, it would > load the Fortran microcode. To run a Cobol application, it would load > the Cobol microcode. This allowed the microcode to be customized to > the particular application. I don't remember if the machine was designed > to be a multitasking machine or not. Still, it's possible that the > microcode could be pagable at task swap time. That was the Burroughs 1700. -Ron Hunsinger ###### From: hnsngr@sirius.com (Ron Hunsinger) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Wed, 14 Oct 1998 17:24:20 -0700 Organization: ErsteSoft Lines: 36 Message-ID: References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> NNTP-Posting-Host: ppp-asok02--062.sirius.net Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: Yet Another NewsWatcher 2.3.1 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news.idt.net!WCG!newshub.sirius.com!newsfiler.sirius.com!hnsngr In article , Neil Franklin wrote: > I was more trying to determine whether the "Fortran in hardware" stories > I have heard were urban legends of the "CISC at the most ridiculous" > variety. Or whether someone had actually designed or even implemented > such a thing. I don't know about Fortran, but I read a report by some guys who did Algol in hardware. They did it using entirely SSI (Small Scale Integration - NAND/NOR/etc) and MSI (Medium Scale Integration - registers and adders and such). No firmware, no microcode, no LSI, nothing like that. The logic was all hard wired. The machine had an object language, but could execute ALGOL source code directly. The original intention was that, as an optimization, you could load a source program into RAM, compile it, save it to disk, and load and run that. But they discovered that direct execution of the source was fast enough that there was no need to take the extra step of compiling. The speed gained by compiling didn't offset the cost of the extra disk write/disk read to save and reload the object code. And, by saving only source, they saved a bunch of disk space. They made it clear that they were executing ALGOL, not tokenized ALGOL or anything like that. The hardware recognized the beginning of a procedure by seeing the bytes 'P', 'R', 'O', 'C', 'E', 'D', 'U', 'R', 'E', ' ' in memory, and did the right thing. This report was not second or third hand. It didn't talk about "some guys in a university somewhere who built this machine that...". It said "*We* built this machine. These are some of the problems *we* ran into. These are some of lessons *we* learned." Their conclusion: "If we had it to do over, we wouldn't do it over." -Ron Hunsinger ###### From: jgd@alpha3.csd.uwm.edu (John G Dobnick) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 14 Oct 1998 21:21:25 GMT Organization: University of Wisconsin - Milwaukee Lines: 39 Message-ID: <7034kl$g4b$1@uwm.edu> References: <702j8t$10m4$1@ausnews.austin.ibm.com> Reply-To: jgd@alpha3.csd.uwm.edu NNTP-Posting-Host: 129.89.7.203 Originator: jgd@alpha3.csd.uwm.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!cpk-news-hub1.bbnplanet.com!chicago-news-feed1.bbnplanet.com!news.bbnplanet.com!uwm.edu!alpha3.csd.uwm.edu!jgd From article <702j8t$10m4$1@ausnews.austin.ibm.com>, by glass2@glass2.cv.lexington.ibm.com: > > I seem to remember one machine, was it the 5500, that had custom microcode > for different languages. Thus, to run a Fortran application, it would > load the Fortran microcode. To run a Cobol application, it would load > the Cobol microcode. This allowed the microcode to be customized to > the particular application. I don't remember if the machine was designed > to be a multitasking machine or not. Still, it's possible that the > microcode could be pagable at task swap time. Right company (Burroughs), wrong machine. You are thinking of the Burroughs 1700 series. Ran an "MCP machine", a "compiler machine", an "ALGOL machine", a "COBOL machine", a FORTRAN machine", and whatever else someone had cooked up in (if I remember properly) S-Language. Dynamically swappable microcode, optimized for the particular task it was running. User writable microstore. An "amorphous" hardware architecture -- the specific microcode running at any given instant "molded" the box into what the user wanted. Neat system. We just couldn't seem to convince the Burroughs sales-critter that we really didn't _need_ lots of (expensive) fast microcode store. We were looking for a machine to _teach_ on, so speed didn't really matter all that much. The sales-critter kept saying "But, it won't run fast without that fast memory". We kept saying "We don't care! It's for _teaching_!!" Burroughs never did sell us one of those boxes. The B5000 was a hard wired, and hard configured, architecture. Very traditional -- except for the stack, and the hardware limits checking, and the virtual memory, and the descriptor (object) implementation, and it's being a multi-processor in the early 1960's (before IBM "invented" all of the above), and... :-) -- John G Dobnick "Knowing how things work is the basis Information & Media Technologies for appreciation, and is thus a University of Wisconsin - Milwaukee source of civilized delight." jgd@csd.uwm.edu ATTnet: (414) 229-5727 -- William Safire ###### Newsgroups: alt.folklore.computers From: ehrice@his.com (Edward Rice) Subject: Re: Fortran source as instruction code (was Re: Rubout) Message-ID: Organization: NDS Date: Thu, 15 Oct 1998 01:29:14 -0400 References: <908089676@ablelink.org> NNTP-Posting-Host: ehrice.his.com Lines: 18 X-Authenticated-User: ehrice Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newspeer.monmouth.com!news.lightlink.com!news4.his.com!user In article , Neil Franklin wrote: > This reminded me of an project (IIRC in the 70s) to build an computer > that directly ran Fortran (II?, 66?, 77?) source as its instruction code. > > Does anyone know any details about this? Was it ever built or was it > just an off beat idea? SYMBOL, a language that looked like Fortran and executed /directly/ on the SYMBOL hardware architecture (which was built in quantity of one or two) came out of Iowa State University in 1969. Rex Rice (no relative) headed the project and delivered the major talks on the subject at the Spring Joint that year. ###### Date: 15 Oct 98 10:36:35 -0800 From: "Charlie Gibbs" Subject: Re: Fortran source as instruction code (was Re: Rubout) References: <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> Message-ID: <1828.592T1768T6364725@sky.bus.com> Newsgroups: alt.folklore.computers Lines: 37 X-Newsreader: THOR 2.5a (Amiga;TCP/IP) NNTP-Posting-Host: news.skybus.com X-Trace: 15 Oct 1998 12:07:03 -0700, news.skybus.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!news.maxwell.syr.edu!news-peer.sprintlink.net!news-backup-east.sprintlink.net!news.sprintlink.net!204.244.4.2!news.westel.com!news.skybus.com!204.244.247.111 In article hnsngr@sirius.com (Ron Hunsinger) writes: >The machine had an object language, but could execute ALGOL source >code directly. The original intention was that, as an optimization, >you could load a source program into RAM, compile it, save it to >disk, and load and run that. But they discovered that direct execution >of the source was fast enough that there was no need to take the extra >step of compiling. The speed gained by compiling didn't offset the >cost of the extra disk write/disk read to save and reload the object >code. And, by saving only source, they saved a bunch of disk space. You can tell that this was done in an academic or research environment. Even though it was faster to directly execute the source code than to compile it and execute the object code (which is pretty impressive), you wouldn't "compile and go" in a production environment. If you're executing the same program hundreds of times, day after day, you're way ahead to compile it just once, because the faster execution of the object code soon pays for itself. As for disk space, the source code would reside on a reel of tape somewhere - the object code that stays on disk would be smaller, so it's a win again. >They made it clear that they were executing ALGOL, not tokenized >ALGOL or anything like that. The hardware recognized the beginning >of a procedure by seeing the bytes 'P', 'R', 'O', 'C', 'E', 'D', >'U', 'R', 'E', ' ' in memory, and did the right thing. That is *scary*. I'm impressed that they got it to work at all. >Their conclusion: "If we had it to do over, we wouldn't do it over." It sounds like one of those things that are fun to try - but only once. -- cgibbs@sky.bus.com (Charlie Gibbs) Remove the first period after the "at" sign to reply. ###### Newsgroups: alt.folklore.computers From: as999@torfree.net (Adrian Boldan) Subject: Re: Fortran source as instruction code (was Re: Rubout) Message-ID: Organization: Toronto Free-Net X-Newsreader: TIN [version 1.2 PL2] References: <908089676@ablelink.org> <6vredi$roc@freenet-news.carleton.ca> <6vrtgh$16es$1@nntp1.u.washington.edu> <6vtj0o$8t8$1@news1.tc.umn.edu> Date: Thu, 15 Oct 1998 20:24:09 GMT Lines: 16 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!news.tele2.nl!newsfeed1.swip.net!swipnet!news.idt.net!newsin.iconnet.net!feed2.nntp.acc.ca!feed.nntp.acc.ca!torfree!as999 George R. Gonzalez (grg@foundsys.com) wrote: : hardware", but these are almost certainly blue-sky exercises-- not likely : anything anybody has actually implemented or marketed. My thesys was a design of an ALU with only NAND gates. In 1967. I still have in my library a book by Yaohan Chu about circuit design. -- *** Learn Esperanto - the International Language! *** One language for all | Unu por chiuj, the second for everyone! | la dua por chiu! ###### From: Donald Fisk Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Mon, 19 Oct 1998 19:21:25 +0100 Organization: British Telecommunications plc Lines: 99 Message-ID: <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> References: <702j8t$10m4$1@ausnews.austin.ibm.com> <7034kl$g4b$1@uwm.edu> Reply-To: fiskd@boat.bt.com NNTP-Posting-Host: pc577f.btlabs.bt.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.05 [en] (Win95; I) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet-feed2!btnet!bt!not-for-mail John G Dobnick wrote: > Right company (Burroughs), wrong machine. You are thinking of the > Burroughs 1700 series. Ran an "MCP machine", a "compiler machine", an > "ALGOL machine", a "COBOL machine", a FORTRAN machine", and whatever > else someone had cooked up in (if I remember properly) S-Language. > Dynamically swappable microcode, optimized for the particular task it > was running. User writable microstore. An "amorphous" hardware > architecture -- the specific microcode running at any given instant > "molded" the box into what the user wanted. There are two different meanings to "microcode". One is yours, which is perhaps better known as "virtual machine", and is what Java has, and the other works like this: "Microcoded Instruction Set After an instruction is fetched from memory, it is used as an index into an array of bit arrays (usually called a nano-ROM). Each array element (a bit array) determines which transmission gates are open during the execute part of the particular instruction's fetch-execute cycle. If a bit is ON, the gate is open, otherwise it is closed. The advantage of having a microcoded instruction set is that it is possible to change the instruction set without redesigning the processor, simply by changing the nano-ROM." What the B1700 (and successors) had were virtual machines. These allowed compiled code to run on very different architectures, e.g. the same compiled COBOL program would run on the B1900 (a 24 bit machine) and the B90 (and 8 bit machine). In other words, exactly the same thing as we have today with Sun's "write once, run anywhere" philosophy for Java. I don't know the details of the B1900 instruction set other than its native word size (see above) and the intention that virtual machines of arbitrary word size could be easily implemented using it. The B1900s were built, and the system programming done, in a plant in Liege in Belgium, and the B90s at Cumbernauld in Scotland. Both machines had an MCP (operating system) written in assembly language, along with the virtual machines for COBOL and MPL (Burrough's answer to C) and later Ada. Everything else on the machine was written in high level languages which after compilation were interpreted on the virtual machines. I am unaware of any implementation of Algol (unless you mean Ada) or Fortran on these machines, though it is possible that someone outside Burroughs developed them. I do know a fair amount about the B90 instruction set, which was microcoded in the other sense. The actual processor architecture was not too bad. (It had several 8 and 16 bit registers, and one 8 byte register that could be used for decimal arithmetic, together with a tiny hardware stack. It ran at 4MHz, with instructions on 8 bit registers taking 1 clock cycle and instructions on 16 bit registers taking 2. Other instructions, e.g. jumps and memory accesses, took up to 5 clock cycles.) The problem from the assembly programmer's point of view was that the instruction set wasn't so much designed as having happened almost by accident, and was horrendously asymmetric, making it virtually impossible to compile into, and somewhat intellectually challenging to program at assembly level. Of course, once it was burnt in and significant amounts of the MCP had been written in it, it could not be changed. > Neat system. We just couldn't seem to convince the Burroughs > sales-critter that we really didn't _need_ lots of (expensive) fast > microcode store. We were looking for a machine to _teach_ on, so speed > didn't really matter all that much. The sales-critter kept saying > "But, it won't run fast without that fast memory". We kept saying "We > don't care! It's for _teaching_!!" Burroughs never did sell us one of > those boxes. Well yes, the B1900 and related machines had the reputation ofrunning like tortoises on valium. I can only assume that you intended that they be used for programming in COBOL, as MPL skills weren't particularly useful outside of a Burroughs environment and nothing much else ran on them. > The B5000 was a hard wired, and hard configured, architecture. Very > traditional -- except for the stack, and the hardware limits checking, > and the virtual memory, and the descriptor (object) implementation, > and it's being a multi-processor in the early 1960's (before IBM > "invented" all of the above), and... :-) Smiley noted. In summary, the Burroughs mainframes had impressivearchitectures, their B1900s had an interesting architecture with disappointing run time performance, and the B90s had the worst instruction set of any machine I've ever programmed. > John G Dobnick "Knowing how things work is the basis -- Le Hibou (mo bheachd fhe/in: my own opinion) "it's just that in C++ and the like, you don't trust _anybody_, and in CLOS you basically trust everybody. the practical result is that thieves and bums use C++ and nice people use CLOS." -- Erik Naggum ###### From: jgd@alpha3.csd.uwm.edu (John G Dobnick) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: 19 Oct 1998 19:34:38 GMT Organization: University of Wisconsin - Milwaukee Lines: 32 Message-ID: <70g48e$vv$1@uwm.edu> References: <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> Reply-To: jgd@alpha3.csd.uwm.edu NNTP-Posting-Host: 129.89.7.203 Originator: jgd@alpha3.csd.uwm.edu Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newspump.sol.net!uwm.edu!alpha3.csd.uwm.edu!jgd From article <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk>, by Donald Fisk : > John G Dobnick wrote: >> Burroughs 1700 series. Ran an "MCP machine", a "compiler machine", an >> "ALGOL machine", a "COBOL machine", a FORTRAN machine", and whatever >> else someone had cooked up in (if I remember properly) S-Language. > > There are two different meanings to "microcode". One is yours, which > is perhaps better known as "virtual machine", and is what Java has, and > the other works like this: True. I believe Burroughs actually used the term "S-machine" to describe these "virtual machines". >> Neat system. We just couldn't seem to convince the Burroughs >> sales-critter that we really didn't _need_ lots of (expensive) fast >> microcode store. > > Well yes, the B1900 and related machines had the reputation ofrunning like tortoises on > valium. I can only assume that you intended > that they be used for programming in COBOL, as MPL skills weren't > particularly useful outside of a Burroughs environment and nothing much > else ran on them. Not _necessarily_ COBOL. Teaching general computer architecture. On something _other_ than a PDP-8 or S/360. Variety, and all that. -- John G Dobnick "Knowing how things work is the basis Information & Media Technologies for appreciation, and is thus a University of Wisconsin - Milwaukee source of civilized delight." jgd@csd.uwm.edu ATTnet: (414) 229-5727 -- William Safire ###### From: jmfbahciv@aol.com Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Tue, 20 Oct 98 09:34:07 GMT Organization: UltraNet Communications, Inc. Lines: 33 Message-ID: <70hpr5$5ot$3@ligarius.ultra.net> References: <702j8t$10m4$1@ausnews.austin.ibm.com> <7034kl$g4b$1@uwm.edu> <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> NNTP-Posting-Host: d4.dial-13.mbo.ma.ultra.net X-Complaints-To: abuse@ultra.net X-Ultra-Time: 20 Oct 1998 10:49:09 GMT X-Newsreader: News Xpress Version 1.0 Beta #4 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!feed1.news.luth.se!luth.se!news-peer-europe.sprintlink.net!news.sprintlink.net!Sprint!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!newsfeed.xcom.net!news.ultranet.com!d4 In article <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk>, Donald Fisk wrote: >John G Dobnick wrote: > >> Right company (Burroughs), wrong machine. You are thinking of the >> Burroughs 1700 series. Ran an "MCP machine", a "compiler machine", an >> "ALGOL machine", a "COBOL machine", a FORTRAN machine", and whatever >> else someone had cooked up in (if I remember properly) S-Language. >> Dynamically swappable microcode, optimized for the particular task it >> was running. User writable microstore. An "amorphous" hardware >> architecture -- the specific microcode running at any given instant >> "molded" the box into what the user wanted. > >There are two different meanings to "microcode". One is yours, which >is perhaps better known as "virtual machine", and is what Java has, and >the other works like this: > > "Microcoded Instruction Set > >After an instruction is fetched from memory, it is used as an >index into an array of bit arrays (usually called a nano-ROM). >Each array element (a bit array) determines which transmission >gates are open during the execute part of the particular >instruction's fetch-execute cycle. If a bit is ON, the gate is >open, otherwise it is closed. Thank you for that explanation. Things are becoming clearer now since I had never considered the virtual machine type as microcode. No wonder the kiddies are confused. /BAH Subtract a hundred and four for e-mail. ###### From: Donald Fisk Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Tue, 20 Oct 1998 18:23:45 +0100 Organization: British Telecommunications plc Lines: 33 Message-ID: <362CC721.902C6021@bt-sys.spamblock.bt.co.uk> References: <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> <70g48e$vv$1@uwm.edu> NNTP-Posting-Host: pc577f.btlabs.bt.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mailer: Mozilla 4.05 [en] (Win95; I) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsfeed.wirehub.nl!btnet-peer!btnet-feed2!btnet!bt!not-for-mail John G Dobnick wrote: > True. I believe Burroughs actually used the term "S-machine" > to describe these "virtual machines". The code was called S-code, and the virtual machines were calledS-Machines, S standing for Soft. Most of you will be aware of, and some of you may have read, a novel of that title by another William S. Burroughs. There is also a rock group named after the novel.) Does anyone know which Burroughs thought of the term Soft Machine first? > Not _necessarily_ COBOL. Teaching general computer architecture. > On something _other_ than a PDP-8 or S/360. Variety, and all that. The two hardware architectures I would have used as examples of gooddesign are the Burroughs mainframes and the PDP-11 (assuming I have the time frame right). Of course, you might also have wanted to expose them to architectures which didn't live up to either of those two ideals, but were more widely used in commercial environments, in which case IBM mainframes would have been a good choice. Just my 0.02 euro's worth. > John G Dobnick -- Le Hibou (mo bheachd fhéin: my own opinion) "it's just that in C++ and the like, you don't trust _anybody_, and in CLOS you basically trust everybody. the practical result is that thieves and bums use C++ and nice people use CLOS." -- Erik Naggum ###### From: spalding@iol.ie (Nick Spalding) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Wed, 21 Oct 1998 12:05:55 +0100 Organization: Ireland On-Line Message-ID: <3634bfbe.83956253@news.iol.ie> References: <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> <70g48e$vv$1@uwm.edu> <362CC721.902C6021@bt-sys.spamblock.bt.co.uk> Reply-To: spalding@iol.ie NNTP-Posting-Host: dialup-0389.dublin.iol.ie Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Newsreader: Forte Agent 1.5/32.451 Lines: 9 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!newsgate.cistron.nl!het.net!news-feed.inet.tele.dk!bofh.vszbr.cz!dispose.news.demon.net!demon!news-lond.gip.net!news.gsl.net!gip.net!iol!iol.ie!not-for-mail Donald Fisk wrote: > some of you may have read, a novel of that title by another William S. > Burroughs. Grandson (I think) of the founder of the Burroughs company. Family money from the business supported him all his life. -- Nick Spalding ###### From: tph@longhorn.uucp (Tom Harrington) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Rubout) Date: 22 Oct 1998 17:58:31 GMT Organization: Mechanist Industries Lines: 23 Message-ID: <70nro7$at22@eccws1.dearborn.ford.com> References: <6vtj0o$8t8$1@news1.tc.umn.edu> <700dce$950$1@irk.zetnet.co.uk> Reply-To: tph@rmi.net NNTP-Posting-Host: cs0053.eld.ford.com X-Newsreader: TIN [version 1.2 PL2] Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-dc.gip.net!news-peer.gip.net!news.gsl.net!gip.net!newsxfer3.itd.umich.edu!jobone!dailyplanet.srl.ford.com!eccws1.dearborn.ford.com!longhorn!tph lisard@zetnet.co.uk wrote: : On 1998-10-12 grg@foundsys.com said: : :You may find references to many PhD theses with titles like "APL in : :hardware", but these are almost certainly blue-sky exercises-- not : :likely anything anybody has actually implemented or marketed. : Of course, there have been a fair few attempts at putting Forth in : hardware, with various degrees of commercial success. Not generally : including the text interpreter, but Forth's intermediate representation : (pure threaded code) does lend itself well to direct-to-hardware : representation. And Sun is apparently still working on Java-in-silicon implementations. This would be for Java bytecodes rather than source code, but it's interesting that the Java virtual machine seems destined to become a real machine. -- Tom Harrington --------- tph@rmii.com -------- http://rainbow.rmii.com/~tph "And you may ask yourself, well, how did I get here?" -Talking Heads Cookie's Revenge: ftp://ftp.rmi.net/pub2/tph/cookie/cookies-revenge.sit.hqx ###### From: lisard@zetnet.co.uk Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Rubout) Date: 22 Oct 1998 21:06:19 GMT Lines: 20 Message-ID: <70o6ob$d32$6@irk.zetnet.co.uk> References: <362CC721.902C6021@bt-sys.spamblock.bt.co.uk> NNTP-Posting-Host: man-090.dialup.zetnet.co.uk X-Trace: irk.zetnet.co.uk 909090379 13410 194.247.41.112 (22 Oct 1998 21:06:19 GMT) NNTP-Posting-Date: 22 Oct 1998 21:06:19 GMT X-Everything: Net-Tamer V 1.08X Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!btnet-peer!btnet-feed1!btnet!peer.news.zetnet.net!zetnet.co.uk!not-for-mail On 1998-10-20 donald.fisk@bt-sys.spamblock.bt.co.uk said: :> True. I believe Burroughs actually used the term "S-machine" :> to describe these "virtual machines". :The code was called S-code, and the virtual machines were :calledS-Machines, S standing for Soft. Most of you will be aware :of, and some of you may have read, a novel of that title by another :William S. Burroughs. There is also a rock group named after the :novel.) Does anyone know which Burroughs thought of the term Soft :Machine first? I believe that William S Burroughs was actually the grandson of the founder of Burroughs Corp. It wouldn't surprise me to discovere that he knew what was happening in the company, although I doubt he was a board member, somehow. ;> -- Communa (together) we remember... we'll see you falling you know soft spoken changes nothing to sing within her... ###### From: amroth@zetnet.co.uk.NOJUNK (Phil Edwards) Newsgroups: alt.folklore.computers Subject: Re: Fortran source as instruction code (was Re: Rubout) Date: Thu, 22 Oct 1998 22:07:13 GMT Message-ID: <363499af.4265200@news.zetnet.co.uk> References: <362B8325.E0D0911D@bt-sys.spamblock.bt.co.uk> <70g48e$vv$1@uwm.edu> <362CC721.902C6021@bt-sys.spamblock.bt.co.uk> <3634bfbe.83956253@news.iol.ie> NNTP-Posting-Host: man-191.dialup.zetnet.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: roch.zetnet.co.uk 909093974 23524 194.247.40.242 (22 Oct 1998 22:06:14 GMT) NNTP-Posting-Date: 22 Oct 1998 22:06:14 GMT X-Newsreader: Forte Agent 1.5/16.451 Lines: 19 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.belnet.be!news-raspail.gip.net!news-lond.gip.net!news.gsl.net!gip.net!dispose.news.demon.net!demon!peer.news.zetnet.net!zetnet.co.uk!not-for-mail spalding@iol.ie (Nick Spalding) wrote: >Donald Fisk wrote: > >> some of you may have read, a novel of that title by another William S. >> Burroughs. > >Grandson (I think) of the founder of the Burroughs company. Family >money from the business supported him all his life. I thought Burroughs would have been the heir to Burroughs Machines, but that his father had sold the company during the Depression. I don't know if there would have been enough money from the sale of the company to support William S. Burroughs III, but I tend to doubt it. Phil -- Phil Edwards http://www.users.zetnet.co.uk/amroth/ "Is there some unsuspected dignity to this foolishness?"