Re: Theodore Adorno, a prophet of data systems design

• Previous message: Richard Heathfield: "Re: Theodore Adorno, a prophet of data systems design"
• In reply to: bob7094: "Re: Theodore Adorno, a prophet of data systems design"
• Next in thread: Randy Howard: "Re: Theodore Adorno, a prophet of data systems design"
• Reply: Randy Howard: "Re: Theodore Adorno, a prophet of data systems design"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: 7 Jan 2004 23:46:48 -0800

"bob7094" <bobdotsevenohninefour@comcast.net> wrote in message news:<UKGdnYR-pZGQ-maiRVn-uA@comcast.com>...
> "Edward G. Nilges" <spinoza1111@yahoo.com> wrote in message
> news:f5dda427.0401060043.6323e350@posting.google.com...
> > "bob7094" <bobdotsevenohninefour@comcast.net> wrote in message
> news:<h-adnYlRdoY5rGeiRVn-sA@comcast.com>...
> > > "Edward G. Nilges" <spinoza1111@yahoo.com> wrote in message
> > > news:f5dda427.0401031327.4e873b10@posting.google.com...
> > > > Theodore Wiesengrund Adorno, who died in 1970, was a prophet of the
> > > > modern organization of information labor. Unlike traditional Marxist
> > > > thinkers, for whom the "proletariat" was traditional, male, industrial
> > > > workers, Adorno saw as an actual information worker (who was engaged,
> > > > when he fled Hitler, in studying radio listening by a think tank
> > > > funded by RCA) the real conditions of real white-collar work.
> > > >
> > > > Here is Adorno's prophetic statement about what REALLY goes on in
> > > > computer work today.
> > > >
>
> <snip>
>
> > Adorno is hard to understand but I am at a loss to how you've
> > interpreted the passage.
>
> I wasn't trying to interpret Adorno's passage, but yours. I find that
> trying to interpret deliberatly obfuscatory text is not worth my time.

Adorno was not "deliberately obfuscatory". Instead, because of
outdated and poor translations of his work issued by Continuum in the
1970s, he gained a reputation with English and American academics for
being "deliberately obfuscatory".

He did have a theory that thought was necessarily "about" a subject
matter that was independent of thought, and that thought needed to be
adequate to the subject matter and this makes him at time seem
"obfuscatory".

>
> > I have described in full my connection with Roosevelt University's
> > 1401, and with Princeton at a later date, below. But here I would say
> > that you and I have witnessed different stages in this process at
> > Princeton.
>
> Ah, that makes sense. Smaller institutions often get obsolete equipment.

Yup. My goal was to make it adequate to growing demands.

> It's far better to have something than nothing at all. I thought that you
> were relating the 1401 and Princeton from your previous posts (although
> I couldn't find it on google, which is why I asked). Sorry about that.
>
>
> <snip>
>
> > My understanding is that during your era, Princeton ran a 360 upgraded
> > to near-supercomputer status, but before that time it probably had a
> > 7094 and these computers often enslaved poor little 1401s to do their
> > input and output. This was the case down at the University of Chicago,
> > which supplied Roosevelt with its only halfway decent assembler for
> > Symbolic Programming System.
> >
> > The 360 (65?) of your era had a channel IO architecture which meant
> > that IO cycles did not tie up the machine as they did on the 7094 so
> > it did not need IBM 1401s. Furthermore, the 1401 was based on a six
> > bit character, and the 360 as you know used an 8 bit character.
> >
> > My discovery at Roosevelt was that IBM had undersold the 1401 to
> > Roosevelt. It was touted to large firms and large universities as a
> > mere IO slave with fast tapes, fast card readers, and an incredibly
> > solid and reliable printer on which you could play music because it
> > was both noisy and precise. The 1401 was touted to midsize firms and
> > smaller schools as an adjunct to tab equipment.
>
> Yes, when I started programming at Princeton (as a high school junior),
> there was a 7044 and a 7094. The brave little 1401 indeed did the

Once upon a time, a brave little 1401 was working for an IBM 7094 at
Princeton University. The IBM 7094 and the IBM 7044 would laugh at the
poor, brave little 1401.

"We can compile Fortran", said the mean computers. "We can play chess
and beat people. What can you do except read, print and punch?'"

So one day, late at night, the brave little 1401 escaped from the
computer center!

The brave little 1401 was frightened as it rolled on its casters down
Prospect Avenue.

"Modify address", it said, "but Prospect avenue sure is scarey late at
night. I sure hope I don't meet the Phantom of Fine Hall."

But the brave little 1401 made it to the Dinky and got the last train
to Princeton Junction. It then used the money it had saved up to get
the Pennsylvania Railroad west to Chicago, where it applied for a job
at Roosevelt University.

There it met a programmer who fixed its Fortran compiler.

The brave little 1401 was very happy.

Today, it is still running in Fiji doing the payroll for an
agricultural cooperative!

> physical labor of reading and punching cards and printing on it's modest
> 1403. The haughty 7094 got all the credit even though all it did was sit
> there and think and move the arms on it's great big disk drive a bit. The
> only demeaning work it ever did was read SYSPIT and write SYSPOT for
> the few lunatic users who tried writing SNOBOL programs. (My,
> this Watty Piper style is hard).
>
> There was a 1410 at PPPL that I only got to use once. If I remember

For listeners, this is plasma physics lab. It was in existence when I
was there.

The 1410 was a souped up 1401 in an effort spearheaded, I believe, by
a John Haanstra, who wanted IBM to build on the 1401 rather than go to
the 360. He felt that the 1401's architecture (no registers, one
memory of six bit bytes, completely variable length numbers and
strings delimited by word marks) could handle all business and most
scientific tasks.

Orlando Wilson, the first supt. of police in Chicago not on the
"take", who took office in the 1950s after a major scandal,
computerized the police department with the IBM 1440 which even had
remote terminals and multiprocessing capability to support early "911"
dispatching, which rousted the cops out of the donut shops. Roosevelt
was offered the 1440 when the cop shop upgraded in 1973 to IBM 370s
and I had sugarplums dancing in my head, for I could then develop my
very own operating system for remote access and win fame, and glory
(this actually happened quite a lot to ordinary slobs in the early
years). I didn't actually know how to write an operating system, but I
figured, what the hell.

However, Roosevelt was talked out of this by IBM. They had severely
punished Haanstra for being too persistent in his views when Watson
Jr. decided that the 360 was the way to go and they wanted to
eliminate old 1401s with the IBM logo.

And IBM was right. The 1401 architecture wasn't equal to the growing
tasks. Variable length operands and "can't add, doesnt try" and its
table lookup algorithm for arithmetic would have seriously retarded
progress. For this reason, I got into BAL and Cobol on small size 360s
including the ridiculous IBM 360/20.

The 1401's architecture does survive in a remote sense in interpreted
languages with weak typing and unrestricted string lengths like perl
and like rexx. It placed the actual machine language as close as
possible to the user's perceived needs.

This was part of a broader push in the 1960 to design the hardware for
a specific language which produced the B5xxx series from Burroughs and
various Fortran and Basic based desk calculators. It died out with
some rather powerful "Basic calculators" used mostly by dorks like me
in the late 1970s. In retrospect we shuddha spent 1000 bucks on an
Apple II, and when I in fact realized that Basic calculators were a
dead end I became an early Macintosh adopter.

> correctly it had a Fortran II compiler (vs the two versions of Fortran IV

Roosevelt's compiler was Fortran II. It had no capability for
manipulating strings, although it had, in the FORMAT statement, a
strikingly complete facility for formatting strings.

> on the 7044 and 7094). I saw/heard the 1403 playing music; you
> had to be careful about not cutting the expensive 14 inch wide ribbon in
> half.

My breakage was as follows.

(1) Once and only once did I use the character in location 200 which
controlled the printing of the memory between 201 and 233 in such a
way as to actually cut a hole.

(2) Occasionally I would forget to insert the valid character and the
printer would eject in a runaway and rather celebratory, Dionysiac
fashion. This wasn't breakage because all you had to do was refold the
paper.

(3) I dropped more than one card deck. However, I'd written a utility
for sequencing all decks produced without exception and specified its
use so I never destroyed data.

>
> > However, I discovered that by clever use of overlays such as I found
> > in the Fortran compiler, you could create programs of arbitrarily
> > large size even on Roosevelt's card-only system. I was able therefore
> > to create a primitive data base management system based on Reverse
> > Polish evaluation of selection and format specs and reasonably
> > elaborate graphical output using the line printer.
>
> Shirley you had 7-track tapes. The biggest problem with overlays is

Nope. No tapes and no disks. 8K. You could smoke in the computer room.
I'm still trying to quit.

> they're SLOW, since everytime you change from one overlay to another
> you have to reload the unshared portions from disk (if you had it) or tape.

If all you have are cards, you insert the data deck behind phase one
and in front of all the other decks. Phase one stays resident in low
memory past the punch area (381 or thereabouts and loads subsequent
phases.

If you carefully managed IO, you could ensure that the process time
was bound not by the CPU but by the card reader time. The goal was to
have the card reader continuously operating at its highest rated
speed.

Now, the 1401 had no "channels" and could not therefore overlap its
CPU time with IO. However, several operations would in fact do
combined IO such as Read and Print and even Read, Print and Punch.
Therefore you did primitive anticipatory "scheduling" so as to
maximise your use of combined operations.

At the time I left I was in fact working on the beginnings of an
"operating system" (the 1401 had no operating system and was a
barenaked machine) which would allow users to call IOCS routines with
requests to read, write or punch which would be scheduled with other
requests.

The problem was NO secondary storage and to implement my idea, I
realized that to simulate secondary storage, the operator would have
to stand all day feeding and retrieving punched cards. The punch would
be punching new cards ALL THE TIME as "pages" were "swapped" to
cards.

This was to me obviously nuts, both for ecological and labor reasons.
I should have explored instead the electronics of interfacing an audio
tape encoded with digital signals to the 1401 but this would have
damaged university property.

You use markers carefully and artistically to show where to insert the
card decks and you draw a big, black, diagonal line down the deck to
provide positive proof that it is always sorted.

You documented everything in a manual.

Above all, you conducted yourself in an anti-authoritarian fashion so
the union operators like you. Since I'd been an SEIU member this was
easy.

>
> <snip>
>
> > Were you the guy who gave me fits trying to maintain a general purpose
> > mailing system written in BAL? Did you write it? No hard feelings if
> > you did. Like my system at Roosevelt it tried to handle a variety of
> > formats but it had a strange bug in that it would occasionally
> > "forget" to clear its memory.
>
> No, I never did any BAL. We used assembler-F (the macro assembler).

As did I, mostly rewrites and fixes since mainframe usage was
declining between 1987 and 1992. Outside of The Office of Population
Research, very little academic work was still being done on the
mainframe. Instead, Sun workstations and the supercomputer then at
John von Neumann were all the rage.

I did develop, mostly on my own time, a Rexx parser based on a parser
generator in Rexx but Rexx usage died at Princeton when Perl came out
in the 1990s. Oh well.

> One summer I wrote a program to process the computer center accounting
> tapes -- a perfect COBOL application -- in assembler.
> I mostly helped develop RJE software for remote access to the 360/91't
> from 360/20's and 1130's located at various places on campus and other
> colleges in New Jersey.
>
Don't know if I ever saw the accounting program.

You were fortunate I think to work at Princeton. A number of people
who were in charge when I was there were there when you were there,
and they were a great bunch of people. They were in fact key to John
Nash's later recovery and this should have been in the film.
> <snip>
>
> -- Bob Mills

• Previous message: Richard Heathfield: "Re: Theodore Adorno, a prophet of data systems design"
• In reply to: bob7094: "Re: Theodore Adorno, a prophet of data systems design"
• Next in thread: Randy Howard: "Re: Theodore Adorno, a prophet of data systems design"
• Reply: Randy Howard: "Re: Theodore Adorno, a prophet of data systems design"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]