29 Apr
2025
29 Apr
'25
6:55 a.m.
From: Tom Lyon <pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; I don't think the code for it is in the
TUHS archives.
Was the Interdata system in use at Bell Labs for actual work once
the port was complete?
ISTR there was a meeting with Interdata about changes in the architecture
that Bell Labs wanted, that Interdata didn't want to make. What
was the full story?
Any other info would be welcome.
Thanks,
Arnold
29 Apr
29 Apr
7:54 a.m.
On Tue, Apr 29, 2025 at 12:55:05AM -0600, arnold(a)skeeve.com wrote:
Tom described some of the C portability problems in
Inter-UNIX Portability
https://archive.org/details/CLanguagePortability_Sept77/page/n15/mode/2up
https://akapugsblog.wordpress.com/wp-content/uploads/2018/05/inter-unix_por…
From: Tom Lyon
<pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; I don't think the code for it is in the
TUHS archives.
Was the Interdata system in use at Bell Labs for actual work once
the port was complete?
ISTR there was a meeting with Interdata about changes in the architecture
that Bell Labs wanted, that Interdata didn't want to make. What
was the full story?
Steve Johnson on the hardware problems
https://www.tuhs.org/pipermail/coff/2023-August/001672.html
Any other info would be welcome.
Thanks,
Arnold
30 Apr
30 Apr
7:08 a.m.
Thanks for this, and to everyone else who replied. I'm guessing
the hardware problems prevented the system from being used further.
Thanks,
Arnold
Jonathan Gray <jsg(a)jsg.id.au> wrote:
On Tue, Apr 29, 2025 at 12:55:05AM -0600,
arnold(a)skeeve.com wrote:
Tom described some of the C portability problems in
Inter-UNIX Portability
https://archive.org/details/CLanguagePortability_Sept77/page/n15/mode/2up
https://akapugsblog.wordpress.com/wp-content/uploads/2018/05/inter-unix_por…
From: Tom
Lyon <pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; I don't think the code for it is in the
TUHS archives.
Was the Interdata system in use at Bell Labs for actual work once
the port was complete?
ISTR there was a meeting with Interdata about changes in the architecture
that Bell Labs wanted, that Interdata didn't want to make. What
was the full story?
Steve Johnson on the hardware problems
https://www.tuhs.org/pipermail/coff/2023-August/001672.html
>
> Any other info would be welcome.
>
> Thanks,
>
> Arnold
> 
29 Apr
29 Apr
5 p.m.
On Tue, Apr 29, 2025 at 12:55 AM <arnold(a)skeeve.com> wrote:
So on the marketing side, the Interdata port wound up being sold in the US
by The Wollogong Group (TWG). They marketed it to Harris computer users
more generally. Tom and Steve were frustrated that they couldn't market this
in .au so effectively sold their rights to TWG who did market it.
And then they branched out, using their V6 license to do a number of other
things:
* Eunice for VMS (4BSD environment for VMS)
* TCP/IP for VMS (the TCP. from 4.2BSD (and later 4.3BSD) ported to VMS,
as extracted from Eunice, not to be confused with TVG's Multinet).
* TCP/IP for a bunch of other System V systems (mostly 3B2*'s) as well as
HP (both HPUX and non unix systems) and a few other early odd-balls
that my memory isn't responding well to.
* Unix internals training with the Lions Book, including AT&T and Wollgong
intellectual property markings (I wish I'd tried hard to snag a copy when
I worked there: I had to settle for an Nth generation copy from a coworker).
* Some other early porting work that may have just been marketing material
to compete with Unisoft / raise funding. I just saw this stuff once at the TWG
offices and was given some hand-wavey explanation that amounted to
"don't worry about testing this"...
* IP/TCP for DOS (though this was an independent thing, done at MIT,
the licensing material for this product, IIRC, included some kind of
unix permission that confused me at the time, but it may have just
been lawyering to CYA rather than including anything).
Warner
From: Tom Lyon <pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; I don't think the code for it is in the
TUHS archives.
Was the Interdata system in use at Bell Labs for actual work once
the port was complete?
ISTR there was a meeting with Interdata about changes in the architecture
that Bell Labs wanted, that Interdata didn't want to make. What
was the full story?
Any other info would be welcome. 
5:18 p.m.
On 4/29/25 10:00 AM, Warner Losh wrote:
And then they branched out, using their V6 license to
do a number of other
things:
* Eunice for VMS (4BSD environment for VMS)
Minor quible - Eunice was developed by David Kashtan at SRI, but
then maintained and marketed the The Wollongong Group.
(I used it on VMS at UCSB in the early 1980s).
Jon
P.S. A non-Unix Interdata memory. At UCSB we used 3 Interdata
machines as terminal switchers. You could connect from a terminal
in certain terminal rooms to any of a number of computers on campus that
had support for the switching protocol. The names of the Interdata
machines were "huey", "duey", and "louie". There was
allegedly a
fourth machine called "kablooey" that was used as a backup.
5:25 p.m.
On Tue, Apr 29, 2025 at 11:18 AM Jon Forrest <nobozo(a)gmail.com> wrote:
On 4/29/25 10:00 AM, Warner Losh wrote:
True, I mentioned it only because they used their V6 license which
had unusually favorable terms to do so... Which is an unanticipated
"long hand" of the interdata port into the late 80s and early 90s.
Warner
And then they branched out, using their V6
license to do a number of other
things:
* Eunice for VMS (4BSD environment for VMS)
Minor quible - Eunice was developed by David Kashtan at SRI, but
then maintained and marketed the The Wollongong Group.
(I used it on VMS at UCSB in the early 1980s). Jon
P.S. A non-Unix Interdata memory. At UCSB we used 3 Interdata
machines as terminal switchers. You could connect from a terminal
in certain terminal rooms to any of a number of computers on campus that
had support for the switching protocol. The names of the Interdata
machines were "huey", "duey", and "louie". There was
allegedly a
fourth machine called "kablooey" that was used as a backup.
5:35 p.m.
On 4/29/25 10:00 AM, Warner Losh wrote:
On Tue, Apr 29, 2025 at 12:55 AM
<arnold(a)skeeve.com> wrote:
>
>> From: Tom Lyon <pugs78(a)gmail.com>
>>
>> I was pleased to learn that the first port of S to UNIX was on the
>> Interdata 8/32, which I had my part in enabling.
>
> I would love to hear more about the Interdata port and what
> happened with it afterwards. Interdata seems to have disappeared
> into the dustbin of history. And Unix on it apparently never
> got out of Bell Labs; I don't think the code for it is in the
> TUHS archives.
>
Two different ports.
TWG's port was written at the University of Wollongong, extensive
documentation and SIMH emulation survives.
No trace of the BTL port exists that I know of
Interdata was bought by Perkin-Elmer, Perkin-Elmer sold it to
Concurrent Computer Corp.
30 Apr
30 Apr
7:21 a.m.
Al Kossow <aek(a)bitsavers.org> wrote:
Two different ports.
TWG's port was written at the University of Wollongong, extensive
documentation and SIMH emulation survives.
Ah! I was shown a V6 port on SIMH, but Bell Labs didn't do
a port of V6. This explains it.
No trace of the BTL port exists that I know of
Interdata was bought by Perkin-Elmer, Perkin-Elmer sold it to
Concurrent Computer Corp.
CCI made the Tahoe that 4.4 ran on, but I'm guessing it's
a different architecture than the Interdata?
Thanks,
Arnold
7:35 a.m.
On Apr 30, 2025, at 12:21 AM, arnold(a)skeeve.com wrote:
Al Kossow <aek(a)bitsavers.org> wrote:
The story of this first port (by Richard Miller) here:
https://www.tuhs.org/Archive/Distributions/Other/Interdata/uow103747.pdf
Two different ports.
TWG's port was written at the University of Wollongong, extensive
documentation and SIMH emulation survives.
Ah! I was shown a V6 port on SIMH, but Bell Labs didn't do
a port of V6. This explains it.
7:39 a.m.
On Apr 30, 2025, at 12:35 AM, Bakul Shah <bakul(a)iitbombay.org> wrote:
On Apr 30, 2025, at 12:21 AM, arnold(a)skeeve.com wrote:
More details in Richard's invited talk @ Usenix:
https://www.usenix.org/legacy/publications/library/proceedings/usenix98/inv…
Al Kossow <aek(a)bitsavers.org> wrote:
The story of this first port (by Richard Miller) here:
https://www.tuhs.org/Archive/Distributions/Other/Interdata/uow103747.pdf Two different ports.
TWG's port was written at the University of Wollongong, extensive
documentation and SIMH emulation survives.
Ah! I was shown a V6 port on SIMH, but Bell Labs didn't do
a port of V6. This explains it. 
29 Apr
29 Apr
5:58 p.m.
On Tue, Apr 29, 2025 at 1:00 PM Warner Losh <imp(a)bsdimp.com> wrote:
* IP/TCP for DOS (though this was an independent
thing, done at MIT,
the licensing material for this product, IIRC, included some kind of
unix permission that confused me at the time, but it may have just
been lawyering to CYA rather than including anything).
The PC/IP software from MIT included a port of the "Portable C Compiler" to
generate 8086-era code. It ran on a Unix machine and built binaries that
you downloaded to the PC. So you need an ATT source license to get the full
PCIP dev kit.
6:25 p.m.
Yes, that was one of the RTS compilers for the NU machine. John Romkey may
have done it, as he was the primary person behind PCIP, but I can not claim
I remember who did the 8086 backend. IIRC Jack Test did the 68K backend.
The RTS crew had the NU machine and NU bus that went with it. Very tean
project Neat project. Similar idea, in fact to what CMU was doing with the
Intel Mutlibus called the distributed front end (we had started with LSI-11
and cost reduced it to 8086 on a Intel Multibus). Andy Bechtolsheim would
take with him to Stanford and rework with a 68K which became the
Stanford Network Terminal - which used the RTS's C compilers. It's all
very mixed up. [ Don't tell me there was not an open source culture back in
the old days by the way].
Anyway the MIT RTS foilks made hardware and PCC back ends for the 68K,
Z8000 and 8086. I believe that each had separate assemblers, tjt who
sometimes reads this list might know more, as he wrote the 68K assembler.
IIRC they had a common linker which is was rewrite/extension to the
original V7 linker or maybe the 4.1 linker.
Anyone with a V7 license could get it. If you had a PC license you get
get the source to Romkey's PCIP. If you did not a license, you could only
get a binary kit.
ᐧ
On Tue, Apr 29, 2025 at 1:59 PM Rich Salz <rich.salz(a)gmail.com> wrote:
On Tue, Apr 29, 2025 at 1:00 PM Warner Losh <imp(a)bsdimp.com> wrote:
* IP/TCP for DOS (though this was an independent
thing, done at MIT,
the licensing material for this product, IIRC, included some kind of
unix permission that confused me at the time, but it may have just
been lawyering to CYA rather than including anything).
The PC/IP software from MIT included a port of the "Portable C Compiler"
to generate 8086-era code. It ran on a Unix machine and built binaries that
you downloaded to the PC. So you need an ATT source license to get the full
PCIP dev kit.
6:38 p.m.
On Tue, Apr 29, 2025, 12:26 PM Clem Cole <clemc(a)ccc.com> wrote:
Yes, that was one of the RTS compilers for the NU
machine. John Romkey
may have done it, as he was the primary person behind PCIP, but I can not
claim I remember who did the 8086 backend. IIRC Jack Test did the 68K
backend. The RTS crew had the NU machine and NU bus that went with it.
Very tean project Neat project. Similar idea, in fact to what CMU was
doing with the Intel Mutlibus called the distributed front end (we had
started with LSI-11 and cost reduced it to 8086 on a Intel Multibus). Andy
Bechtolsheim would take with him to Stanford and rework with a 68K which
became the Stanford Network Terminal - which used the RTS's C compilers.
It's all very mixed up. [ Don't tell me there was not an open source
culture back in the old days by the way].
Anyway the MIT RTS foilks made hardware and PCC back ends for the 68K,
Z8000 and 8086. I believe that each had separate assemblers, tjt who
sometimes reads this list might know more, as he wrote the 68K assembler.
IIRC they had a common linker which is was rewrite/extension to the
original V7 linker or maybe the 4.1 linker.
Anyone with a V7 license could get it. If you had a PC license you get
get the source to Romkey's PCIP. If you did not a license, you could only
get a binary kit.
ᐧ
On Tue, Apr 29, 2025 at 1:59 PM Rich Salz <rich.salz(a)gmail.com> wrote:
Yea. The MIT compiler was also shipped with Venix/86
Warner
On Tue, Apr 29, 2025 at 1:00 PM Warner Losh <imp(a)bsdimp.com> wrote:
* IP/TCP for DOS (though this was an independent
thing, done at MIT,
the licensing material for this product, IIRC, included some kind of
unix permission that confused me at the time, but it may have just
been lawyering to CYA rather than including anything).
The PC/IP software from MIT included a port of the "Portable C Compiler"
to generate 8086-era code. It ran on a Unix machine and built binaries that
you downloaded to the PC. So you need an ATT source license to get the full
PCIP dev kit.
6:46 p.m.
On 4/29/25 11:25 AM, Clem Cole wrote:
Yes, that was one of the RTS compilers for the NU
machine. John Romkey may have done it, as he was the primary person behind PCIP, but I
can not claim I remember who did the 8086 backend. IIRC Jack Test did the 68K backend.
The RTS crew had the NU machine and NU bus that
went with it. Very tean project Neat project. Similar idea, in fact to what CMU was
doing with the Intel Mutlibus called the distributed
front end (we had started with LSI-11 and cost reduced it to 8086 on a Intel Multibus).
Andy Bechtolsheim would take with him to Stanford
and rework with a 68K which became the Stanford Network Terminal - which used the
RTS's C compilers. It's all very mixed up. [ Don't tell
me there was not an open source culture back in the old days by the way].
Anyway the MIT RTS foilks made hardware and PCC back ends for the 68K, Z8000 and 8086. I
believe that each had separate assemblers, tjt who
sometimes reads this list might know more, as he wrote the 68K assembler. IIRC they had
a common linker which is was rewrite/extension to
the original V7 linker or maybe the 4.1 linker.
Anyone with a V7 license could get it. If you had a PC license you get get the source
to Romkey's PCIP. If you did not a license, you
could only get a binary kit.
I have an image of the "MIT Compiler Tape" with a bunch of different PCC ports
from a couple of different institutions.
On the Stanford side, SUMACC was a hack of the compiler to work with the Macintosh
I had the source for it at one point, but haven't been able to find it on any of my
backups.
7:07 p.m.
On 4/29/25 11:46 AM, Al Kossow wrote:
I have an image of the "MIT Compiler Tape"
with a bunch of different PCC ports from a couple of different institutions.
I forgot I put this up a while ago
http://bitsavers.org/bits/MIT/trix/MIT_Compiler_Tape
I wish I had more information about what went on wrt Steve Ward's lab
7:10 p.m.
On Tue, Apr 29, 2025 at 1:07 PM Al Kossow <aek(a)bitsavers.org> wrote:
On 4/29/25 11:46 AM, Al Kossow wrote:
I've tried building that for modern systems and the code is definitely
very K&R...
In one spot, IIRC, there's even a vax queue instruction.
Warner
I have an image of the "MIT Compiler
Tape" with a bunch of different PCC ports from a couple of different institutions.
I forgot I put this up a while ago
http://bitsavers.org/bits/MIT/trix/MIT_Compiler_Tape
I wish I had more information about what went on wrt Steve Ward's lab
7:45 p.m.
Thanks for the refresh -- National chip (16032), not the Z8000 but right
idea.
ᐧ
On Tue, Apr 29, 2025 at 3:07 PM Al Kossow <aek(a)bitsavers.org> wrote:
On 4/29/25 11:46 AM, Al Kossow wrote:
I have an image of the "MIT Compiler
Tape" with a bunch of different PCC
ports from a couple of different
institutions.
I forgot I put this up a while ago
http://bitsavers.org/bits/MIT/trix/MIT_Compiler_Tape
I wish I had more information about what went on wrt Steve Ward's lab
8:14 p.m.
On Tuesday, April 29th, 2025 at 12:45 PM, Clem Cole <clemc(a)ccc.com> wrote:
Thanks for the refresh -- National chip (16032), not
the Z8000 but right idea.
ᐧ
On Tue, Apr 29, 2025 at 3:07 PM Al Kossow <aek(a)bitsavers.org> wrote:
> On 4/29/25 11:46 AM, Al Kossow wrote:
>
> > I have an image of the "MIT Compiler Tape" with a bunch of different
PCC ports from a couple of different institutions.
>
> I forgot I put this up a while ago
> http://bitsavers.org/bits/MIT/trix/MIT_Compiler_Tape
>
> I wish I had more information about what went on wrt Steve Ward's lab
Just fyi, thread from a couple years back in which Steve Johnson offered some
recollections of his time involved in the Interdata 8/32 work at BTL:
https://tuhs.org/mailman3/hyperkitty/list/tuhs@tuhs.org/thread/WH4IEMFZEZED…
His recollections above touch on the microcode bug that prevented the Interdata port from
being viable beyond porting exercises.
- Matt G.
11:27 p.m.
On Tue, Apr 29, 2025 at 11:46:19AM -0700, Al Kossow wrote:
On 4/29/25 11:25 AM, Clem Cole wrote:
in the TUHS archive
Distributions/UCB/4.3BSD/new.tar.gz
contains sumacc.tar.Z which has source
Yes, that was one of the RTS compilers for the NU
machine. John Romkey
may have done it, as he was the primary person behind PCIP, but I can
not claim I remember who did the 8086 backend. IIRC Jack Test did the
68K backend. The RTS crew had the NU machine and NU bus that went with
it. Very tean project Neat project. Similar idea, in fact to what CMU
was doing with the Intel Mutlibus called the distributed front end (we
had started with LSI-11 and cost reduced it to 8086 on a Intel
Multibus). Andy Bechtolsheim would take with him to Stanford and rework
with a 68K which became the Stanford Network Terminal - which used the
RTS's C compilers. It's all very mixed up. [ Don't tell me there was
not an open source culture back in the old days by the way].
Anyway the MIT RTS foilks made hardware and PCC back ends for the 68K,
Z8000 and 8086. I believe that each had separate assemblers, tjt who
sometimes reads this list might know more, as he wrote the 68K
assembler. IIRC they had a common linker which is was rewrite/extension
to the original V7 linker or maybe the 4.1 linker.
Anyone with a V7 license could get it. If you had a PC license you get
get the source to Romkey's PCIP. If you did not a license, you could
only get a binary kit.
I have an image of the "MIT Compiler Tape" with a bunch of different PCC ports
from a couple of different institutions.
On the Stanford side, SUMACC was a hack of the compiler to work with the Macintosh
I had the source for it at one point, but haven't been able to find it on any of my
backups. 
6:28 p.m.
On 4/29/25 1:00 PM, Warner Losh wrote:
And then they branched out, using their V6 license to
do a number of other
things:
* Eunice for VMS (4BSD environment for VMS)
"Congratulations. You aren't running Eunice."
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU chet(a)case.edu http://tiswww.cwru.edu/~chet/
9:03 p.m.
On Tue, 29 Apr 2025 at 14:36, Chet Ramey via TUHS <tuhs(a)tuhs.org> wrote:
On 4/29/25 1:00 PM, Warner Losh wrote:
Perl has some good quips hidden in Configure but that was always one of my
favorites, even well before I had any real idea what Eunice was.
Personally I want a t-shirt that says "configure: error: cannot guess host
type; you must specify one" which nicely sums up my experiences with both
historic computing and my general life.
-Henry
And then they branched out, using their V6
license to do a number of
other
things:
* Eunice for VMS (4BSD environment for VMS)
"Congratulations. You aren't running Eunice." 
30 Apr
30 Apr
9:17 p.m.
At Tue, 29 Apr 2025 00:55:05 -0600, arnold(a)skeeve.com wrote:
Subject: [TUHS] Any Interdata war stories?
In about 1981 or 1982 the chemistry (IIRC) department at University of
Calgary had an Interdata 8/32 that was, at least for some time around
then, running Unix. I remember poking around on it briefly, but I don't
remember much more than that about it.
--
Greg A. Woods <gwoods(a)acm.org>
Kelowna, BC +1 250 762-7675 RoboHack <woods(a)robohack.ca>
Planix, Inc. <woods(a)planix.com> Avoncote Farms <woods(a)avoncote.ca>
From: Tom Lyon <pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; 
1 May
1 May
4:44 a.m.
I found this USENIX paper from Steve Johnson that has a lot of detail about
the
Interdata port (postscript format):
"C and the AT&T Unix Port--A Personal History"
https://www.usenix.org/legacy/publications/library/proceedings/usenix98/inv…
On Mon, Apr 28, 2025 at 11:55 PM <arnold(a)skeeve.com> wrote:
From: Tom Lyon
<pugs78(a)gmail.com>
I was pleased to learn that the first port of S to UNIX was on the
Interdata 8/32, which I had my part in enabling.
I would love to hear more about the Interdata port and what
happened with it afterwards. Interdata seems to have disappeared
into the dustbin of history. And Unix on it apparently never
got out of Bell Labs; I don't think the code for it is in the
TUHS archives.
Was the Interdata system in use at Bell Labs for actual work once
the port was complete?
ISTR there was a meeting with Interdata about changes in the architecture
that Bell Labs wanted, that Interdata didn't want to make. What
was the full story?
Any other info would be welcome.
Thanks,
Arnold
20
days inactive
22
days old
22 comments
13 participants
participants (13)
-
Al Kossow -
arnold＠skeeve.com -
Bakul Shah -
Chet Ramey -
Clem Cole -
Greg A. Woods -
Henry Bent -
Jon Forrest -
Jonathan Gray -
Rich Salz -
segaloco -
Tom Lyon -
Warner Losh