15 Mar
2024
15 Mar
'24
9:45 a.m.
In another thread there's been some discussion of Coherent. I just came
across this very detailed history, just posted last month. There's much
more to it than I knew.
https://www.abortretry.fail/p/the-mark-williams-company
Marc
15 Mar
15 Mar
11:16 a.m.
Interesting little history about Coherent. They were
one of a few companies building UNIX-like systems
from scratch without using UNIX source code in the
early 1980's. Robert Schwartz represented the Mark
Williams Company on the /usr/group standards
effort resulting in the /usr/group Standard in 1984.
Robert was very insistent that members of the
/usr/group standards group did not have to be
UNIX source licensees.
Heinz
On 3/14/2024 8:45 PM, Marc Rochkind wrote:
In another thread there's been some discussion of
Coherent. I just
came across this very detailed history, just posted last month.
There's much more to it than I knew.
https://www.abortretry.fail/p/the-mark-williams-company
Marc
1 p.m.
Another detail. There was lawyerly concern about the code being stolen, and
we (127) were asked to find ways to test, absent their source, whether they
had just stolen our source and built the binaries. It was soon concluded
that there were enough details different to definitively say that at least
most of the work was done in a clean room, as advertised, but the piece I
liked best is that their PPT(1) program (ASCII art showing a paper tape
rendering the argument text) did not include the original, and just
discovered, bug that mispunched, if I remember right, the letter 'R'.
-rob
On Fri, Mar 15, 2024 at 4:16 PM Heinz Lycklama <heinz(a)osta.com> wrote:
Interesting little history about Coherent. They were
one of a few companies building UNIX-like systems
from scratch without using UNIX source code in the
early 1980's. Robert Schwartz represented the Mark
Williams Company on the /usr/group standards
effort resulting in the /usr/group Standard in 1984.
Robert was very insistent that members of the
/usr/group standards group did not have to be
UNIX source licensees.
Heinz
On 3/14/2024 8:45 PM, Marc Rochkind wrote:
In another thread there's been some
discussion of Coherent. I just
came across this very detailed history, just posted last month.
There's much more to it than I knew.
https://www.abortretry.fail/p/the-mark-williams-company
Marc
7:03 p.m.
On Fri, Mar 15, 2024 at 3:00 AM Rob Pike <robpike(a)gmail.com> wrote:
Another detail. There was lawyerly concern about the
code being stolen, and we (127) were asked to find ways to test, absent their source,
whether they had just stolen our source and built the binaries. It was soon concluded that
there were enough details different to definitively say that at least most of the work was
done in a clean room, as advertised, but the piece I liked best is that their PPT(1)
program (ASCII art showing a paper tape rendering the argument text) did not include the
original, and just discovered, bug that mispunched, if I remember right, the letter
'R'.
Along those lines, Dennis Ritchie wrote up a summary of the event on
USENET; apparently in 1998 (I had no idea it was this late):
https://groups.google.com/g/alt.folklore.computers/c/_ZaYeY46eb4/m/5B41Uym6…
COHERENT (version 4) was my introduction to Unix (or Unix-like)
systems. I bought it from an ad in the back of "Computer Shopper" or
one of those things; my first inkling that it was rather different
from actual Unix was that the `lc` command they had picked up
(probably from York or Toronto) was not present on SunOS or 4.3BSD.
Similarly, the manual was rather different: it didn't have the usual
sectioned Unix manual, but rather an alphabetical "Lexicon" and
chapters discussing specific topics (editors, UUCP, etc); in
retrospect I thought their manual and its format was rather nice; it
was certainly well-written and beautifully typeset. Regardless of
that, I pretty quickly left COHERENT behind for NetBSD.
COHERENT was an early casualty of Linux's success, and I don't think
it ever occupied much more than a niche, but it was an interesting
system. I've booted it a few times under emulation out of nostalgia.
I had a very small hand in the opening of their sources. I knew that
Stephen Ness had archived copies, at the request of Bob Swartz, and I
wrote to him about the system overall and ended with something like,
"if those sources are available, I'd love to see them." He responded
that due to my message, he'd corresponded with Swartz, who had agreed
to release the sources under the 3 clause BSD license.
Incidentally, Robert Swartz was the father of the late Aaron Swartz.
- Dan C.
On Fri, Mar 15, 2024 at 4:16 PM Heinz Lycklama
<heinz(a)osta.com> wrote:
>
> Interesting little history about Coherent. They were
> one of a few companies building UNIX-like systems
> from scratch without using UNIX source code in the
> early 1980's. Robert Schwartz represented the Mark
> Williams Company on the /usr/group standards
> effort resulting in the /usr/group Standard in 1984.
> Robert was very insistent that members of the
> /usr/group standards group did not have to be
> UNIX source licensees.
>
> Heinz
>
> On 3/14/2024 8:45 PM, Marc Rochkind wrote:
> > In another thread there's been some discussion of Coherent. I just
> > came across this very detailed history, just posted last month.
> > There's much more to it than I knew.
> >
> > https://www.abortretry.fail/p/the-mark-williams-company
> >
> > Marc
> >
>
7:43 p.m.
On Fri, Mar 15, 2024 at 9:03 AM Dan Cross <crossd(a)gmail.com> wrote:
On Fri, Mar 15, 2024 at 3:00 AM Rob Pike <robpike(a)gmail.com> wrote:
Sorry, just to clarify: I meant I had no idea Dennis's posting about
the event happened so late; by 1998 USENET was basically overrun by
spam. Obviously, the inspection trip had happened much earlier.
- Dan C.
Another detail. There was lawyerly concern about
the code being stolen, and we (127) were asked to find ways to test, absent their source,
whether they had just stolen our source and built the binaries. It was soon concluded that
there were enough details different to definitively say that at least most of the work was
done in a clean room, as advertised, but the piece I liked best is that their PPT(1)
program (ASCII art showing a paper tape rendering the argument text) did not include the
original, and just discovered, bug that mispunched, if I remember right, the letter
'R'.
Along those lines, Dennis Ritchie wrote up a summary of the event on
USENET; apparently in 1998 (I had no idea it was this late):
https://groups.google.com/g/alt.folklore.computers/c/_ZaYeY46eb4/m/5B41Uym6…
8:10 p.m.
I have already posted this in another thread (on non-BTL C compilers), but
it's more relevant here. My 1985 review of Coherent for BYTE Magazine:
https://www.mrochkind.com/mrochkind/docs/Byte-Pick-Coherent-Theos.pdf
I see that I went into some detail. For example:
"Of the 77 requests in the Version 7 nroff. only 31 are present in Coherent
(the most useful 31. however)."
And this, although I'm sure there were incompatibilities I didn't uncover:
"Coherent has all the Version 7 system calls except nice (which sets a
process's priority). and they seem to be used in the same way. It should be
easy to port C programs between Coherent and UNIX Version 7."
On the whole my review was very positive.
Marc
On Fri, Mar 15, 2024 at 7:43 AM Dan Cross <crossd(a)gmail.com> wrote:
On Fri, Mar 15, 2024 at 9:03 AM Dan Cross
<crossd(a)gmail.com> wrote:
--
*My new email address is mrochkind(a)gmail.com <mrochkind(a)gmail.com>*
On Fri, Mar 15, 2024 at 3:00 AM Rob Pike <robpike(a)gmail.com> wrote:
> Another detail. There was lawyerly concern about the code being
stolen, and
we (127) were asked to find ways to test, absent their source,
whether they had just stolen our source and built the binaries. It was soon
concluded that there were enough details different to definitively say that
at least most of the work was done in a clean room, as advertised, but the
piece I liked best is that their PPT(1) program (ASCII art showing a paper
tape rendering the argument text) did not include the original, and just
discovered, bug that mispunched, if I remember right, the letter 'R'.
Along those lines, Dennis Ritchie wrote up a summary of the event on
USENET; apparently in 1998 (I had no idea it was this late):
https://groups.google.com/g/alt.folklore.computers/c/_ZaYeY46eb4/m/5B41Uym6…
Sorry, just to clarify: I meant I had no idea Dennis's posting about
the event happened so late; by 1998 USENET was basically overrun by
spam. Obviously, the inspection trip had happened much earlier.
- Dan C.
9:42 p.m.
On Fri, Mar 15, 2024 at 08:10:31AM -0600, Marc Rochkind wrote:
I have already posted this in another thread (on
non-BTL C compilers), but
it's more relevant here. My 1985 review of Coherent for BYTE Magazine:
https://www.mrochkind.com/mrochkind/docs/Byte-Pick-Coherent-Theos.pdf
It's interesting that it says that $495 for Coherent was a good deal.
(That's over $1700 in 2024 dollars.) On the other hand, in 1984, the
New York Times reported that IBM had "cut the price of the PC/XT" to
$2,520 for a machine with 256k RAM, a single disk drive, and a
monochrome display. (That's over $7400 in 2024 dollars.)
It's amazing how much hardware and software has gotten cheaper in the
past four decades!
- Ted
8:40 p.m.
On Fri, Mar 15, 2024 at 3:00 AM Rob Pike <robpike(a)gmail.com> wrote:
Another detail. There was lawyerly concern about the
code being stolen
There were similar issues at DEC when we ported VAX Fortran to Ultrix.
Especially
with the port of the VMS linker, the piece of the project I
worked on. We had a member of the Ultrix team writing the code to convert
VMS object file debug information to a.out STABs, but other than that the
team doing the port stayed clear of the Ultrix sources.
-Paul W.
282
days inactive
282
days old
7 comments
6 participants
participants (6)
-
Dan Cross -
Heinz Lycklama -
Marc Rochkind -
Paul Winalski -
Rob Pike -
Theodore Ts'o