... or the proceedings that it's in.
The paper is by Chris Torek entitled "A New Framework for Device Support in
Berkeley Unix" from Proceedings of the UKUUG, London, Summer 1990.
The Google hits I'm getting in the proceedings suggest I'd like a copy of
the full thing.
Closest I've found is from 2005 or 2006 on archive.org... Nothing in the
TUHS archives I was able to find....
This paper is referenced in Chris Torek's "Device Configuration in 4.4BSD"
which only ever seemed to circulate in draft form. That I have a pdf of
which I converted from a ps that was on NetBSD.org...
Any chance I can get a copy of it? Or will I need to figure out
inter-library loan again for the first time in almost 2 decades...
Warner
I know this is a strange place to ask, but it was suggested to me that some people who may know may follow this list...
Anyone on here used IBM's XLC in very old versions?
Anyone know what the argument -qdebug=austlib does?
I can't seem to find any documentation that says... It would have been an argument for the compiler shipping with AIX 3.2.5, I believe.
Thanks in advance!
Nemo Nusquam:
In this informal survey, I side with Dave, though I prefer to read in my
comfy well-lit chair with tea/coffee/cocoa.B (A very similar thread was
aired on MO last year.)
=====
I should point out that, having at various times spilled hot
chocolate on a tablet and on a paper book, it is much simpler
to recover when it's a tablet.
And a cat can flip pages for you with either technology.
Norman Wilson
Toronto ON
(Curled up on the couch with my laptop, cat just left)
> On Fri, Apr 9, 2021 at 11:34 PM Ed Bradford <egbegb2 at gmail.com <https://minnie.tuhs.org/cgi-bin/mailman/listinfo/tuhs>> wrote:
>
> > Why did a Ph.D., an academic, and a computer scientist not know about UNIX
> > in 1974 or so? 1976? In 1976, some (many?) universities had source code.
> >
>
> Some knowns/givens at the time ...
> 1.) He was a language/compiler type person -- he had created PL/M and that
> was really what he was originally trying to show off. As I understand it
> and has been reported in other interviews, originally CP/M was an attempt
> to show off what you could do with PL/M.
> 2.) The 8080/Z80 S-100 style machines we quite limited, they had very
> little memory, no MMU, and extremely limited storage in the 8" floppies
> 3.) He was familiar with RT/11 and DOS-11, many Universities had it on
> smaller PDP-11s as they ran on an 11/20 without an MMU also with limited
> memory, and often used simple (primarily tape) storage (DECtape and
> Cassette's) as the default 'laboratory' system, replacing the earlier PDP-8
> for the same job which primarily ran DOS-8 in those settings.
> 4.) Fifth and Sixth Edition of Unix was $150 for university but to run it,
> it took a larger at least 11/40 or 45, with a minimum of 64Kbytes to boot
> and really need the full 256Kbytes to run acceptably and the cost of a 2.5M
> byte RK05 disk was much greater per byte than tape -- thus the base system
> it took to run it was at least $60K (in 1975 dollars) and typically cost
> about two to four times that in practice. Remember the cost of
> acquisition of the HW dominated many (most) choices.
>
> *I**'ll take a guess, but it is only that.* I *suspect* he saw the S-100
> system as closer to a PDP-11/20 'lab' system than as a small
> timesharing machine. He set out with CP/M to duplication the functionality
> from RT/11. He even the naming of the commands was the same as what DEC
> used (*e.g.* PIP) and used the basic DEC style command syntax and parsing
> rules.
That is about it. CP/M predates the Altair / S-100 bus, and was designed for a heavily hacked Intellec-8 system.
CP/M was developed on a PDP-10 based 8080 simulator in 1974. It was developed for the dual purposes of creating a “native” PL/M compiler and to create the “astrology machine”.
The first versions of CP/M were written (mostly) in PL/M. To some extent, in 1974 both Unix and CP/M were research systems, with a kernel coded in a portable language — but aimed at very different levels of hardware capability.
In 1975 customers started to show up and paid serious money for CP/M (Omron, IMSAI) - from that point on the course for Kildall / DRI was set.
The story is here: https://computerhistory.org/blog/in-his-own-words-gary-kildall/?key=in-his-… <https://computerhistory.org/blog/in-his-own-words-gary-kildall/?key=in-his-…>
> I wonder. IBM introduced the IBM PC in August of 1981.
> That was years after a non-memory managed version of
> Unix was created by Heinze Lycklama, LSX. Is anyone
> on this list familiar with Bell Labs management thoughts
> on selling IBM on LSX rather than "dos"?
IBM famously failed to buy the well-established CP/M in
1980. (CP/M had been introduced in 1974, before the
advent of the LSI-11 on which LSX ran.) By then IBM had
settled on Basic and Intel. I do not believe they ever
considered Unix and DEC, nor that AT&T considered
selling to IBM. (AT&T had--fortunately--long since been
rebuffed in an attempt to sell to DEC.)
Doug
I'd totally subscribe to your newsletter :P
that's cool, there is a tape dump of the old stuff on bitsavers... the
UniSoft port I think was the original stuff before Bill showed up?
http://bitsavers.trailing-edge.com/bits/Sun/UniSoft_1.3/
along with some ROM images
http://bitsavers.trailing-edge.com/bits/Sun/sun1/
but more pictures and whatnot are always interesting!
-----Original Message-----
From: Earl Baugh
To: Clem Cole
Cc: tuhs(a)minnie.tuhs.org
Sent: 4/10/21 4:02 AM
Subject: Re: [TUHS] SUN (Stanford University Network) was PC Unix
I’ve done a fair amount of research on Sun 1’s since I have one ( and it
has one of the original 68k motherboards with the original proms ).
It’s on my list to create a Sun 1 registry along the lines of the Apple
1 registry. ( https://www.apple1registry.com/
<https://www.apple1registry.com/> )
Right now, I can positively identify 24 machines that still exist. Odd
serial numbering makes it very hard to know exactly how many they made.
Cisco was sued by Stanford over the Sun 1. From what I read, they made
off with some Stanford property ( SW and HW ). Wikipedia mentions this (
and I have some supporting documents as well ). They ended up licensing
from Stanford as part of the settlement. From what I’ve gathered VLSI
licensed the design from Stanford not Andy directly. However the only
produced a few machines and Andy wasn’t all that happy with that. That
was one of the impetus is to getting sun formed and licensing the same
design. I also believe another company ( or 2 )licensed the design but
either didn’t produce any or very very few machines.
You can tell a difference between VLSI boards and the Sun Microsystems
boards because the SUN is all capitalized on the VLSI boards ( and is
Sun on the others ). At least on the few I’ve seen pictures of.
The design was also licensed to SGI — I’ve seen a prototype SGI board
that’s the same thing with a larger PCB to allow some extensions.
And the original CPU boards didn’t have an MMU. They could only run Sun
OS up to 0.9, I believe was the version. When Bill Joy got there, again
from what I’ve gathered, he wanted to bring more of the BSD code over
and they had to change the system board. This is why you see the Sun
1/150 model number ( as an upgrade to the original Sun 1/100 designation
). The rack mounted Sun 1/120 was changed to the 1/170. The same
upgraded CPU board was used in the Sun 2/120 at least initially.
The original Sun OS wasn’t BSD based. It was a V32 variant I believe.
And the original CPU boards were returned to Sun, I believe as part of
the upgrade from the 1/100 to the 1/150. ( Given people had just paid
$10,000 for a machine having to replace the entire machine would’ve been
bad from a customer perspective). Sun did board upgrade trade ups after
this ( I worked at a company and we purchased an upgrade to upgrade a
Sun 3/140 to a Sun 3/110 — the upgrade consisted of a CPU board swap and
a different badge for the box )
Sun then, from when I can tell, sold the original CPU boards to a German
company that was producing a V32 system. They changed out the PROMs but
you can see the Sun logo and part numbers on the boards
I could go on and on about this topic ?
A Sun 1 was a “bucket list” machine for me - and I am still happy that
some friends were willing to take a 17 hour road trip from Atlanta to
Minnesota to pick mine up. ?
After unparking the drive heads it booted up, first try ( I was only
willing to try that without a bunch of testing work because I have some
spare power supplies and a couple plastic tubs of multi bus boards that
came with it ?)
Earl
Sent from my iPhone
On Apr 9, 2021, at 11:13 AM, Clem Cole <clemc(a)ccc.com> wrote:
?
On Fri, Apr 9, 2021 at 10:10 AM Tom Lyon < pugs(a)ieee.org
<mailto:pugs@ieee.org> > wrote:
Prior to Sun, Andy had a company called VLSI Technology, Inc. which
licensed SUN designs to 5-10 companies, including Forward Technology and
CoData, IIRC. The SUN IPR effectively belonged to Andy, but I don't
know what kind of legal arrangement he had with Stanford. But the
design was not generally public, and relied on CAD tools only extant on
the Stanford PDP-10. Cisco did start with the SUN-1 processor, though
whether they got it from Andy or direct from Stanford is not known to
me. When Cisco started (1984), the Sun-1 was long dead already at Sun.
Bits passing in the night -- this very much is what I remember,
expereinced.
<https://mailfoogae.appspot.com/t?sender=aY2xlbWNAY2NjLmNvbQ%3D%3D&type=
zerocontent&guid=57eccb88-2f68-40ed-9f5a-ce8913f2b4cc> ?
Is there any solid info on the Stanford SUN boards? I just know the SUN-1
was based around them, but they aren't the same thing? And apparently cisco
used them as well but 'borrowed' someone's RTOS design as the basis for IOS?
There was some lawsuit and Stanford got cisco network gear for years for
free but they couldn't take stock for some reason?
I see more and more of these CP/M SBC's on ebay/online and it seems odd that
there is no 'DIY' SUN boards... Or were they not all that open, hence why
they kind of disappeared?
-----Original Message-----
From: Jon Steinhart
To: tuhs(a)minnie.tuhs.org
Sent: 4/8/21 7:04 AM
Subject: Re: [TUHS] PC Unix
Larry McVoy writes:
> On Thu, Apr 08, 2021 at 12:18:04AM +0200, Thomas Paulsen wrote:
> > >From: John Gilmore <gnu(a)toad.com>
> > >Sun was making 68000-based systems in 1981, before the IBM PC was
created.
> >
> > Sun was founded on February 24, 1982. The Sun-1 was launched in May
1982.
> >
> > https://en.wikipedia.org/wiki/Sun_Microsystems
> > https://en.wikipedia.org/wiki/Sun-1
>
> John may be sort of right, I bet avb was building 68k machines at
> Stanford before SUN was founded. Sun stood for Stanford University
> Network I believe.
>
> --lm
Larry is correct. I remember visiting a friend of mind, Gary Newman,
who was working at Lucasfilm in '81. He showed me a bunch of stuff
that they were doing on Stanford University Network boards.
Full disclosure, it was Gary and Paul Rubinfeld who ended up at DEC
and I believe was the architect for the microVax who told me about
the explorer scout post at BTL which is how I met Heinz.
Jon
> From: Jason Stevens
> apparently cisco used them as well but 'borrowed' someone's RTOS design
> as the basis for IOS? There was some lawsuit and Stanford got cisco
> network gear for years for free but they couldn't take stock for some
> reason?
I don't know the whole story, but there was some kind of scandal; I vaguely
recall stories about 'missing' tapes being 'found' under the machine room
raised floor...
The base software for the Cisco multi-protocol router was code done by William
(Bill) Yeager at Stanford (it handled IP and PUP); I have a vgue memory that
his initially ran on PDP-11's, like mine. (I think their use of that code was
part of the scandal, but I've forgotten the details.)
> From: Tom Lyon
> the design ... relied on CAD tools only extant on the Stanford PDP-10.
Sounds like SUDS?
Noel
> I developed LSX at Bell Labs in Murray Hill NJ in the 1974-1975
> timeframe.
> An existing C compiler made it possible without too much effort. The
> UNIX
> source was available to Universities by then. I also developed Mini-UNIX
> for the PDP11/10 (also no memory protection) in the 1976 timeframe.
> This source code was also made available to Universities, but the source
> code for LSX was not.
>
> Peter Weiner, the founder of INTERACTIVE Systems Corp.(ISC) in June
> 1977,
> the first commercial company to license UNIX source from Western
> Electric for $20,000. Binary licenses were available at the same time.
> I joined ISC in May of 1978 when ISC was the first company to offer
> UNIX support services to third parties. There was never any talk about
> licensing UNIX source code from Western Electric (WE) from the founding
> of ISC to when the Intel 8086 micro became available in 1981.
> DEC never really targeted the PC market with the LSI-11 micro,
> and WE never made it easy to license binary copies of the UNIX
> source code, So LSX never really caught on in the commercial market.
> ISC was in the business of porting the UNIX source code to other
> computers, micro to mainframe, as new computer architectures
> were developed.
>
> Heinz
The Wikipedia page for ISC has the following paragraphs:
"Although observers in the early 1980s expected that IBM would choose Microsoft Xenix or a version from AT&T Corporation as the Unix for its microcomputer, PC/IX was the first Unix implementation for the IBM PC XT available directly from IBM. According to Bob Blake, the PC/IX product manager for IBM, their "primary objective was to make a credible Unix system - [...] not try to 'IBM-ize' the product. PC-IX is System III Unix." PC/IX was not, however, the first Unix port to the XT: Venix/86 preceded PC/IX by about a year, although it was based on the older Version 7 Unix.
The main addition to PC/IX was the INed screen editor from ISC. INed offered multiple windows and context-sensitive help, paragraph justification and margin changes, although it was not a fully fledged word processor. PC/IX omitted the System III FORTRAN compiler and the tar file archiver, and did not add BSD tools like vi or the C shell. One reason for not porting these was that in PC/IX, individual applications were limited to a single segment of 64 kB of RAM.
To achieve good filesystem performance, PC/IX addressed the XT hard drive directly, rather than doing this through the BIOS, which gave it a significant speed advantage compared to MS-DOS. Because of the lack of true memory protection in the 8088 chips, IBM only sold single-user licenses for PC/IX.
The PC/IX distribution came on 19 floppy disks and was accompanied by a 1,800-page manual. Installed, PC/IX took approximately 4.5 MB of disk space. An editorial by Bill Machrone in PC Magazine at the time of PC/IX's launch flagged the $900 price as a show stopper given its lack of compatibility with MS-DOS applications. PC/IX was not a commercial success although BYTE in August 1984 described it as "a complete, usable single-user implementation that does what can be done with the 8088", noting that PC/IX on the PC outperformed Venix on the PDP-11/23.”
It seems like Venix/86 came out in Spring 1983 and PC/IX in Spring 1984. I guess by then RAM had become cheap enough that running in 64KB of core was no longer a requirement and LSX and MX did not make sense anymore. Does that sound right?
