On 5/6/2021 12:14 AM, Chuck Guzis via cctalk wrote:
On 5/5/21 6:40 PM, Jay Jaeger via cctalk wrote:
While I was in grad school at U.W. (the one in Wisconsin) we had
obtained via surplus an IBM 7094 II from military surplus - I seem to
recall WSMR (White Sands Missle Range). Of course, it had an MG. We
got it put together and with a replaced transistor here and there (one
of which was modern silicon) we got it running. We had NO peripherals,
but my friend Paul came up with the idea to use a sense switch to talk
RS-232 at 9600bps, and using that he got the Purdue University Fast
FORTRAN Translator (PUFFT)loaded and up and running, and I wrote some
support code on a Datacraft 6024 to send it card images and receive...
I remember PUFFT--wasn't that Saul Rosen's baby? Purdue, IIRC, had a
couple of 7094s and a CDC 6500 and a 1401 operated as SPOOL in the
basement of the Math building. Other departments had their own
systems--Industrial Management, for example, had an 1130 that nobody
seemed to be interested in and was available to anyone who wanted to use
it. I think Administration had an S/360 installation. I believe that
there was the odd 1620 scattered here and there.
Apparently, yes: https://dl.acm.org/doi/10.1145/365660.365671
https://docs.lib.purdue.edu/dissertations/AAI7003878/
https://minds.wisconsin.edu/bitstream/handle/1793/57766/TR160.pdf?sequence=1&isAllowed=y
The last two papers involved Edouard Desautels, who was at UW when I was
there - I really like the guy. He coordinated the "systems lab" and I
think was central to acquiring the 7094 II from surplus. The last paper
(which you can actually download unlike Purdue who seems to have locked
their papers up) hints at why that might have been done.
UW had a Univac 1108 and two B5500's when I started (the second loaned
because Burroughs was late delivering the B8500 - and then lied to UW
when the faculty paid a visit - by lashing up a pair of B6500's,
according to one prof. I had who was there.) The 1108 and B5500's were
replaced by a Univac 1110 before I graduated in 1975.
There was also a CDC 3600 and a CDC 1604. The CDC 3600 was damaged in
the infamous Sterling Hall bombing, but I think the 1604 came out fine.
I expect those were both gone by 1975.
The system lab had a PDP-11/20 and a Datacraft 6024, then after UNIX
showed up (V6 and then V7) got a PDP-11/45 and a PDP-11/40 that was not
used for UNIX, and then a PDP-11/70 for UNIX for the CS department
shortly after I graduated and left.
The 7094 II and the Oscar Mayer 1410 were not used for anything serious
and only lasted a couple of years - there were not even any peripherals
for them. I think Prof. Desautels intended they be used for entry level
instructional classes in FORTRAN and COBOL, maybe (see above).
Engineering had a couple of IBM 1620's (one with a disk) that became
disused when they obtained a Datacraft 6024 around 1970, I think. The
electrical engineering department had a PDP-11/20 almost identical to
the one the the CS dept. had that I did some independent study on.
The School of Business, where I worked as student help / grad project
assistant had an IBM 1410 that had been used by the UW Registrar. When
the administration consolidated, they obtained some model of IBM S/360.
There was an old old IBM 7094 with oil core (or maybe even a 7090, I
don't recall) on the 9th floor of the Chemistry building used by one
particular researcher. It had Potter peripherals. At one point it
failed pretty badly, and three of us went over to see if we could figure
out the problem. They ended up calling in a couple of guys from Purdue
to help figure it out.
The Social Science Research Institute had a S/360 (a model 30 or 50, I
think), as well, right across the street from the School of Business.
To try and get their arms around instructional computing, the UW
computing committee (which somehow ended up with an acronym something
like acc-acc) commissioned a DEC PDP-11/70 for students to use,
replacing the Business School 1410 and some other smaller systems. It
was a disaster - it just couldn't keep up.
I recall Rosen's numerical methods course--on the first day, he came in
lugging one of those Monroe or Friden mechanical calculating machines to
demonstrate significance and loss thereof.
PUFFT was one of those language translators that was so permissive (it
would even suggest correct spellings of misspelled keywords, correct
syntax, etc.) Basically a student compiler that would guarantee that
just about any garbage could be turned into a program.
We were able to play with it because it was core-resident - didn't need
any peripherals to run once it got loaded (see the above third paper.)
--Chuck
