steve@polyslo.CalPoly.EDU (Steve DeJarnett) (12/03/88)
With the recent discussions about old GE systems and where GECOS came from, I started to wonder if anyone still used Multics on any systems. Are there any sites out there that run Multics?? Is there any source code available for it (since I'm asking). I presume it would have been written in something like FORTRAN. We hypothesized that maybe Bell Labs or MIT might be. Well, just curious. Thanks in advance, ------------------------------------------------------------------------------- | Steve DeJarnett | Smart Mailers -> steve@polyslo.CalPoly.EDU | | Computer Systems Lab | Dumb Mailers -> ..!ucbvax!voder!polyslo!steve | | Cal Poly State Univ. |------------------------------------------------| | San Luis Obispo, CA 93407 | BITNET = Because Idiots Type NETwork | -------------------------------------------------------------------------------
farber@linc.cis.upenn.edu (David Farber) (12/04/88)
In article <6392@polyslo.CalPoly.EDU> steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: > >available for it (since I'm asking). I presume it would have been written >in something like FORTRAN. > Multics was NOT written in Fortran. It was written in PL/1 (initially according to the NPL spec then the full PL/1 spec). It was probably the first PL/1 compiler written afterthe Share/IBM 3x3 spec. Dave David Farber; Prof. of CIS and EE, U of Penn, Philadelphia, PA 19104-6389 Tele: 215-898-9508; FAX: 215-274-8192 "The fundamental principle of science, the definition almost, is this: the sole test of the validity of any idea is experiment." -- R. P. Feynman
gwyn@smoke.BRL.MIL (Doug Gwyn ) (12/04/88)
In article <6392@polyslo.CalPoly.EDU> steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: >Are there any sites out there that run Multics?? Is there any source code >available for it (since I'm asking). I presume it would have been written >in something like FORTRAN. I think MULTICS is still in use and available (for a stiff fee) from Honeywell. It was written primarily in PL/I.
freedman@cpsc.ucalgary.ca (Daniel Freedman) (12/04/88)
In article <6392@polyslo.CalPoly.EDU>, steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: > > Are there any sites out there that run Multics?? Is there any source code > available for it (since I'm asking). I presume it would have been written > in something like FORTRAN. Well, we have a 5 processor Multics system which we use as a general purpose mainframe. We also do a lot of Multics maintenance and development work for Honeywell (I guess it must be Honeywell Bull these days). Multics itself was written in PL/I except for those bits written in ALM, Multics' assembly language. The funny thing about Multics is that although it is 20 years old, there are oodles of features present on Multics that you won't find anywhere else. Even if you do find them elsewhere, they often aren't as good. I remember being particularly impressed by the whole organization of the system. With all operating systems, one asks one's self "at what level do things stop looking clean and get hacky"? With Multics, it's clean all the way down to the hardware. Incidentally, anyone thinking of writing an emacs should look at Multics emacs. It's the best that I've ever seen (if you are on the net Bernie Greenberg, thankyou). If you have specific questions, feel free to mail them to me, and I'll do my best to answer them, although it's been a little while since I used Multics to write any programs. Dan Freedman University of Calgary Computer Science Department 2500 University Drive N.W. freedman@cpsc.UCalgary.CA Calgary, Alberta, T2N 1N4 ...!{ubc-vision,ihnp4}!alberta!calgary!freedman
rcsmith@anagld.UUCP (Ray Smith) (12/05/88)
In article <6392@polyslo.CalPoly.EDU> steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: > > With the recent discussions about old GE systems and where GECOS came >from, I started to wonder if anyone still used Multics on any systems. Are >there any sites out there that run Multics?? Yep! There are a couple of systems that I am aware of within the DoD. They love the multi-level security. >Is there any source code available for it (since I'm asking). All of the source was available on the systems I worked on. > I presume it would have been written in something like FORTRAN. Nah! As someone else pointed out, it was written primarily in PL/1. Also as someone else said, there are lots of neat features available in Multics. In addition to B2 level security there are many tools to help developers. If you want more details, let me know and I'll dust off my manuals. The main problem with MULTICS appears to be two-fold. 1) It is pretty much tied to the hardware and 2) Honeywell couldn't market its way out of a paper bag. -Ray P.S. Does anyone know of something in the UNIX world similiar to MULTICS' "forum"??? -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Ray Smith | ...!uunet!mimsy!aplcen!\ Analytics, Inc. | ...!netsys!---anagld!rcsmith Suite 200 | ...!ethos! / 9891 Broken Land Parkway | ...!sundc!/ Columbia, MD 21046 | Voice: (301) 381-4300 Fax: (301) 381-5173 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
usenet@cps3xx.UUCP (Usenet file owner) (12/05/88)
in article <6392@polyslo.CalPoly.EDU>, steve@polyslo.CalPoly.EDU (Steve DeJarnett) says:
$ With the recent discussions about old GE systems and where GECOS came
$ from, I started to wonder if anyone still used Multics on any systems. Are
$ there any sites out there that run Multics??
Some Honeywell-Bull systems still run Multics or GCOS. Where they are I
don't know for sure, but my last job was working on Honeywell terminal
emulators and I know some of our customers were using them.
John H. Lawitzke UUCP: ...rutgers!mailrus!frith!fciiho!jhl
Michigan Farm Bureau ...decvax!purdue!mailrus!frith!fciiho!jhl
Insurance Group ...uunet!frith!jhl
"What?!? Real computing at an insurance company?!? AND in Michigan!?!"barmar@think.COM (Barry Margolin) (12/06/88)
In article <6392@polyslo.CalPoly.EDU> steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: > With the recent discussions about old GE systems and where GECOS came >from, I started to wonder if anyone still used Multics on any systems. Are >there any sites out there that run Multics?? Is there any source code >available for it (since I'm asking). I presume it would have been written >in something like FORTRAN. > > We hypothesized that maybe Bell Labs or MIT might be. Honeywell Bull still has about 50 Multics customers, and a total of about 70 systems (there are a few internal systems, and some customers have more than one). MIT discontinued its Multics service last winter. As far as I know, Bell Labs NEVER had a Multics system of their own (they got out of the project pretty early). There are still five Multics systems at the Pentagon, several each at Ford Motor Company and Electronic Data Systems (the computer subsidiary of GM, which also runs GM's internal computer systems), and about 40 systems in Europe (mostly in France). Officially, Multics development has been capped. However, there is still a small group in Honeywell Bull doing support for the existing customers. ("Small" is relative, since there's never been more than about 120 people in the Honeywell Multics development organization.) There's also a group at the University of Calgary's Advanced Computing Technology Centre doing contract work (Honeywell committed to this before Multics was capped). About 95% of the source code of Multics is in PL/I. A Multics system comes with full source code. I don't know of any other way to get a significant portion of it. I'm not sure why you'd want it, either. Not that it's bad, but what makes Multics so good is not the actual code, but the design. The code is not very portable, either; there's lots of implementation-dependent code and specialized Multics extensions, and there aren't many other full PL/I compilers out there, anyway. Barry Margolin Thinking Machines Corp. barmar@think.com {uunet,harvard}!think!barmar
jpd@usl-pc.usl.edu (DugalJP) (12/06/88)
In article <9062@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>) writes: >>Are there any sites out there that run Multics?? Is there any source code > >I think MULTICS is still in use and available (for a stiff fee) from >Honeywell. It was written primarily in PL/I. Multics is indeed still in use. Our two-processor L68/80 Multics is over 13 years old now, but stuck at release MR10.2 forever. Honeywell Bull dropped plans to build new faster hardware to support Multics, and was involved with the University of Calgary in investigating moving key systems software to Unix, as a migration path. I sure would love to have vfile_ running under Unix! -- James -- -- James Dugal, N5KNX USENET: ...!{dalsqnt,killer}!usl!jpd Associate Director Internet: jpd@usl.edu Computing Center US Mail: PO Box 42770 Lafayette, LA 70504 University of Southwestern LA. Tel. 318-231-6417 U.S.A.
webber@athos.rutgers.edu (Bob Webber) (12/06/88)
In article <281@cs-spool.calgary.UUCP>, freedman@cpsc.ucalgary.ca (Daniel Freedman) writes:
< Well, we have a 5 processor Multics system which we use as a general
< purpose mainframe. We also do a lot of Multics maintenance and development
< work for Honeywell (I guess it must be Honeywell Bull these days).
< Multics itself was written in PL/I except for those bits written in
< ALM, Multics' assembly language. The funny thing about Multics is
< that although it is 20 years old, there are oodles of features present
< on Multics that you won't find anywhere else. Even if you do find
< them elsewhere, they often aren't as good.
Is there anything one should be aware of that isn't presented in the
Organick book?
--- BOB (webber@athos.rutgers.edu ; rutgers!athos.rutgers.edu!webber)mac3n@babbage.acc.virginia.edu (Alex Colvin) (12/06/88)
just a gnit to pick ... Multics PL/I wasn't full PL/I, either IBM or ANSI. Effectively, Multics defined the compileable subset of PL/I. This is responsible for the success (such as it is) of PL/I outside IBM.
car@pte.UUCP (Chris Rende) (12/06/88)
In article <32902@think.UUCP>, barmar@think.COM (Barry Margolin) writes: > In article <6392@polyslo.CalPoly.EDU> steve@polyslo.CalPoly.EDU (Steve DeJarnett) writes: > > With the recent discussions about old GE systems and where GECOS came > >from, I started to wonder if anyone still used Multics on any systems. Are > >there any sites out there that run Multics?? > > Honeywell Bull still has about 50 Multics customers, and a total of > about 70 systems > > There are still five Multics systems at the Pentagon, several each at > Ford Motor Company and Electronic Data Systems (the computer > subsidiary of GM, which also runs GM's internal computer systems), and > about 40 systems in Europe (mostly in France). Some of my info maybe a bit out of date since I left EDS over a year ago... At the time I left EDS, they had 3 Multics systems located in Detroit. EDS inherited them from GM. They were all medium to large sized systems running GM applications. As an aside, EDS also had 2 simular mainframes running the DTSS operating system. (DTSS Inc. has nothing to do with Honeywell). Managing the DTSS systems was my job. The university I attended, Oakland University in Rochester MI, has a small Multics system which they still use. I still have an active Multics ID at Oakland University; I login to see what's new now and then. Ford Motor in Dearborn MI also has 3 Multics systems. Other Multics sites that I've heard about include the University of Maryland, University of Texas, a few sites in France, and Honeywell's central Multics system called "System M". Multics is based on a 36-bit architecture. Also, Multics requires some special hardware to be added to Honeywell's computers in order to provide that top of the line security. This hardware is called an "Appending Unit". DTSS and GECOS runs on Honeywell's computers without an Appending Unit. I'd be glad to (try to) answer any questions I can about Multics or DTSS. car. -- Christopher A. Rende Multics,DTSS,Shortwave,Scanners,StarTrek uunet!{umix,edsews}!rphroy!pte!car Minix,PC/XT,TRS-80 Model I: Buy Sell Trade Motorola VME1131 M68020 SVR2 Precise Technology & Electronics, Inc.
barmar@think.COM (Barry Margolin) (12/07/88)
In article <Dec.5.23.16.06.1988.13632@athos.rutgers.edu> webber@athos.rutgers.edu (Bob Webber) writes: >Is there anything one should be aware of that isn't presented in the >Organick book? I don't recall how much the book talks about details of the file system implementation, but it was extensively revised in the early 70's. Additionally, at the time the book was written, the ring security facility was implemented primarily in software (resulting in many unavoidable security holes), and it was later moved into the hardware; all the stuff in the book about the gatekeeper routine is obsolete. Most of the rest of the book is still pretty accurate. The dynamic linker and the the memory mapped file system haven't changed much. Peripheral I/O has been improved quite a bit, but I doubt that the book goes into that much. Barry Margolin Thinking Machines Corp. barmar@think.com {uunet,harvard}!think!barmar
barmar@think.COM (Barry Margolin) (12/07/88)
In article <438@babbage.acc.virginia.edu> mac3n@babbage.acc.virginia.edu (Alex Colvin) writes: >Multics PL/I wasn't full PL/I, either IBM or ANSI. Effectively, Multics >defined the compileable subset of PL/I. This is responsible for the >success (such as it is) of PL/I outside IBM. The language Multics was originally implemented in was a PL/I subset (called EPL), but a full ANSI PL/I compiler was later written, around 1968 or 1969. It was revised extensively (to take advantage of new hardware features) in 1972. Barry Margolin Thinking Machines Corp. barmar@think.com {uunet,harvard}!think!barmar
davidsen@steinmetz.ge.com (William E. Davidsen Jr) (12/10/88)
In article <33003@think.UUCP> barmar@kulla.think.com.UUCP (Barry Margolin) writes: | I don't recall how much the book talks about details of the file | system implementation, but it was extensively revised in the early | 70's. Additionally, at the time the book was written, the ring | security facility was implemented primarily in software (resulting in | many unavoidable security holes), and it was later moved into the | hardware; all the stuff in the book about the gatekeeper routine is | obsolete. This has been discussed in arch a bit, and I have to feel that if someone were going to market an o/s which is already B2 security certified they would buy the Multics rights and put it on the Intel 386. Not to get into a CPU war, but the 386 has just the rings of protection in hardware needed to implement Multics. When I first saw the chip spec I thought of Multics. Being able to do the memory mapped file i/o in hardware via segments is certainly not hard to take, either. Maybe someday we could have home Multics systems? -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me
jpdres10@usl-pc.usl.edu (Green Eric Lee) (12/11/88)
In message <33004@think.UUCP>, barmar@think.COM (Barry Margolin) says: >In article <438@babbage.acc.virginia.edu> mac3n@babbage.acc.virginia.edu (Alex Colvin) writes: >>Multics PL/I wasn't full PL/I, either IBM or ANSI. Effectively, Multics >>defined the compileable subset of PL/I. This is responsible for the >>success (such as it is) of PL/I outside IBM. > >The language Multics was originally implemented in was a PL/I subset >(called EPL), but a full ANSI PL/I compiler was later written, around >1968 or 1969. It was revised extensively (to take advantage of new >hardware features) in 1972. Hmm... Last time I used Multics and PL/1 was maybe '82 or so, but I do seem to recall that Multics PL/1 still was a "subset" of the full ANSI standard at that time (and I don't suspect they've made much change since). Digging out my old Multics PL/1 manual doesn't give much enlightenment, though. Page 1, section 1, INTRODUCTION: "Multics PL/1 is closely related to ANSI PL/1. [...] This manual does not specify all differences between Multics PL/1 and Standard PL/1 [...] see Appendix A of the PL/1 Language Specification" (another manual which I do not, of course, have hanging around cluttering up my bookshelves). -- Eric Lee Green P.O. Box 92191, Lafayette, LA 70509 {ames,mit-eddie,osu-cis,...}!killer!elg, killer!usl!elg, etc.
munck@linus.UUCP (Robert Munck) (12/14/88)
In article <12774@steinmetz.ge.com> davidsen@crdos1.UUCP (bill davidsen) writes: >In article <6360@killer.DALLAS.TX.US> elg@killer.DALLAS.TX.US (Eric Green) writes: > >| Re: 68030 vs. 80386 -- what I'd love to see on the '386 would be a >| Multics-like system. The hardware looks so Multics-like... > > I thought of Multics when I first saw the 386. Come on someone, >Multics is certified B2 secure, and it's written in a high level >language. Can't someone see the market for Multic/386? Could it really >be harder to port than UNIX? ----------------- Well, doing all of Multics turned out to be a bit much for a one-man project, but I am writing a smaller system with the same basic architecture -- persistent segments that can be attached and detached from the address space. I'm into the debugging of something I think can be called an "open-architecture secure kernel." The "open-architecture" comes from the fact that it should be very easy to add device drivers, servers (such as X-Windows), and make enhancements to the system. The "secure" part is because the base system has all the mechanism needed to be certified at B2 level or higher, in the opinions of the security people here at MITRE (and they help do actual certifications). There's conflict, of course, between "open" and "secure," with most of the security becoming impossible if a user-written, untrusted server is started. We're planning to put some effort into wide distribution of the source code of the system later this year. It's been done with government money, from the Ada Joint Program Office, and so is Public Domain. Currently written in Professional Pascal/386 from MetaWare, but we're considering recoding in Ada now that there may be usable compilers for the 386. The kernel is to form the base system for a prototype implementation of DOD-STD-1838A, the CAIS (an OS interface for programming tools). I wouldn't mind some help debugging the trickier stuff like dynamic paging, might even be able to pay for a summer student or two or subsidize a good semester project. Source code on request. -- Bob Munck, MITRE Corporation -- munck@mitre.org, munck@mitre.arpa -- ..!linus!munck -- 617/271-3671
shan@hall.cray.com (Sharan Kalwani) (12/24/88)
In article <1235@cps3xx.UUCP> usenet@cps3xx.UUCP (Usenet file owner) writes: >in article <6392@polyslo.CalPoly.EDU>, steve@polyslo.CalPoly.EDU (Steve DeJarnett) says: >$ I started to wonder if anyone still used Multics on any systems. >Some Honeywell-Bull systems still run Multics or GCOS. I seem to recall that Oakland University in Rochester MI still uses a Honeywell running MULTICS as their major computing engine. >John H. Lawitzke UUCP: ...rutgers!mailrus!frith!fciiho!jhl Hi John! --shan -- "I used to think I was indecisive, but now I'm not so sure.....;-)" Sharan Kalwani, CRI, 400 Town Center Dr. Suite 305, Dearborn MI 48126 Phone: (313) 271-0177 (313) 956-2809 (313) 252-8086 INTERNET: shan@cray.com UUCP: ...!uunet!cray.com!shan
freedman@cpsc.ucalgary.ca (Daniel Freedman) (12/29/88)
By the way, Multics Emacs is the only program I have ever seen that is (mostly at least) documented in Latin! Dan Freedman University of Calgary Computer Science Department 2500 University Drive N.W. freedman@cpsc.UCalgary.CA Calgary, Alberta, T2N 1N4 ...!alberta!calgary!freedman
martin@lakesys.UUCP (Martin Wiedmeyer) (12/30/88)
When by brother (light truck engineer) came home for Christmas from Detroit, he told me that Ford (his employer) was using MULTICS and was just now thinking about a replacement...... Marty Wiedmeyer martin@lakesys.lakesys.COM -- Marty Wiedmeyer martin@lakesys.lakesys.COM