Rick
I believe they used straight-up Dartmouth BASIC, but maybe that's obvious
and does not need to be stated. I have a paper tape exercise saved by
someone who took intro training in use of the system, with the intro
brochure materials, etc. When I printed the paper tape it contained BASIC
code and the output. The first time I printed the tape it was upside down,
with confusing results!
Sorry I dont have the actual BASIC but it very well may be a simH GE mini
tape file(s) out there GE 225 or 235. I seem to remember seeing this but
did not find after a quick google search just now.
Bill
On Sat, Oct 26, 2024, 1:35 PM Rick Bensene via cctalk <cctalk(a)classiccmp.org>
wrote:
Hello, all,
I know that there has been great effort to gather up and make available
via emulation (SIMH) timeshared Operating Systems for DEC machines, as well
as Hewlett Packard 2000-Series Timeshared BASIC systems, but I was
wondering if there has been any efforts made to archive and perhaps emulate
any versions of General Electric's timesharing systems?
GE's architecture was similar to Hewlett Packard's timeshared BASIC
(2000A, C, C', F, /ACCESS) in that there was a main computer that took care
of managing the user-space and running user programs, and a communications
processor that acted as the front-end that did the terminal services
handling.
At one point, GE's timesharing system service was the largest timeshared
computing service bureau out there, getting an early start in timesharing
out of the timesharing research done at Dartmouth, as GE computers were
used in this research. The GE timeshare service had local dial-up lines in
most major cities, and eventually were connected into Tymnet, further
adding to the places where a local dial-up number could get you into any of
a number of different GE timeshare systems that were connected to the
network.
Has anyone done work on emulating any of GE's processors (200-series,
400-series, DATANET machines) that were used in the Mark I and Mark II
timesharing systems?
GE's place in timesharing history is quite significant, and seems
certainly deserving of efforts to procure and preserve the code, and
perhaps make it live on through emulation. However, GE being the huge
entity that it is (and was back then), getting hands on the code as well as
permission to do anything with it could be a challenge that may have been
attempted and failed who knows how many times over the years.
I'm just wondering if anyone out there may have old listings, mag-tapes,
or card decks laying around that have the source(or binary distributions)
for any of these GE timesharing systems?
If, so, any such media should certainly be put in the hands of an entity
that can assure that they are preserved, and perhaps at some point, made
available online so that others who may have interest could begin work on
emulating these systems.
I thought of this today because an old memory came to the fore out of the
blue. The memory is very clear despite the many years that have elapsed
since then. It is as clear in my mind as it was the day it happened. I
have no idea why it has stuck so clearly in my mind.
Back in high school, there was a service man that would come in to
repair/tune-up the Teletype model 33-ASR's we used to dial into the school
district's HP Timeshared BASIC system. These machines were heavily used by
students who didn't always treat them gently, and the 33-ASR's weren't
really designed for the use they were exposed to, so he would come in
frequently to fix machines that had broken down.
After he had worked on a Teletype, he would dial-in to some kind of
timeshare system with a local phone number and run some test programs to
validate the proper operation of the terminal. and then log his work by
running a BASIC program that would ask him questions about the job, and
he'd fill in the answers.
One day, I happened to be working at a Teletype next to the one he was
working on, and he had just finished up his work on the machine. He and I
had chatted numerous times in the past, so he was comfortable with me, as I
was with him. I watched as he dialed up the phone number for the system
and I memorized it as he dialed it. I also watched as he entered his
account to log into the system. It was IBB00999, and the password was
"INFO". I couldn't see the print out from my angle, so I had to watch his
fingers as he typed in the information. I tried to be as inconspicuous as
possible while watching him, but he made no attempt to block my view or
otherwise keep me from seeing what he was doing. He always made a point of
taking the printout of his session with him rather than leaving it on the
machine or tearing it off and tossing it in the trash. I figure that was a
security measure as the at least the user ID would be listed on the paper.
But, it didn't stop someone from watching him enter the information. I
had become pretty good at watching people's fingers on the keyboard to
figure out what they were typing.
After he had finished his work, I dialed up the system, and when it asked
"USER NUMBER--", I typed in "IBB00999,INFO", and pressed [RETURN].
I
guess I had observed what he'd typed correctly, because then the TTY
clattered out "SYSTEM--", which I did not happen to note the service
man's
answer to this query, but I figured it wanted to know what language to use,
so I typed in "BASIC", and hit [RETURN]. The system then said "NEW OR
OLD--", and since I didn't know what programs were already in the
directory, I typed in NEW, and then pressed [RETURN]. I was then greeted
with "READY". I'd seen this NEW/OLD used on a DEC RSTS/E system, so I
knew
it meant either to start fresh with a new program, or if OLD was typed,
it'd want to know the name of a program to load from the catalog.
I typed in a simple BASIC program, something like generating a listing of
numbers and their square roots from 1 to 100, and typed "RUN", and it
paused for a short time, then began rattling off the list of numbers. I
got nervous, though, and logged out after the program finished (I didn't
know how to stop it, though I later figured out that pressing the [BREAK]
key would work, just like it did on the HP Timeshared BASIC system.
Over the following days, I logged into the system and played with it here
and there, making sure to only log short sessions so as not to rack up too
much time, as the company that the service guy worked for may have been
charged for the online time. I never stayed logged in for more than
perhaps 10 minutes at a time, and the programs I tried didn't chew up much
in the way of resources. I didn't save any programs, nor did I try to do
anything that would leave traces of my visits, other than the fact that I
had dialed in and logged in.
The system seemed to have a FORTRAN subsystem also, and I was able to
enter a small FORTRAN program and RUN it, which was kind of cool. I had
learned FORTRAN by taking a class at a local community college that had an
IBM 360/30, so I knew enough to be able to enter a simple FORTRAN program
and try it out. I thought it was really cool that this system could do
both BASIC and FORTRAN. I tried to see if it had COBOL (which I had also
learned by taking another class at a community college), but alas, it did
not.
I don't know if the languages on the system were interpreted,
semi-compiled (into pseudo-code which was then interpreted), or fully
compiled into machine code. I do know that when you typed RUN to execute a
program, the system would pause for some time before execution began, so
there must have been some kind of processing going on that could have been
compilation of some form.
I remember that the system seemed to be quite a bit more responsive than
the HP 2000C Timeshared BASIC system the school had access to. It would
generally respond to commands immediately, while sometimes the HP system
would pause a bit before anything happened. Admittedly, I knew that our HP
2000C system was very busy all the time during the school days as it was
shared by quite a few different school districts in the county. I had no
idea how many users may have been on the GE timeshared system.
It would take longer on the GE system after you typed RUN on a BASIC
program before the program started executing, but once it did, it would rip
through the program considerably faster than the HP system, especially when
it came to doing lots of math. I once entered a program that would
calculate factorials, and I ran the programs side by side, one on the HP
2000C, and the other on the GE timeshared system (I didn't know it was a GE
system at the time, but figured it out later). I gave the GE system a
little head-start because of the processing before the program ran. The
program would generate a list of integers and their factorials from 1 to
40. As the numbers got larger, the HP timeshared system would print out
the integer and some spaces, then stop for a bit, taking longer as the
factorial was computed, and then print the factorial. The program was the
same on both systems, but the GE system never paused during the output, and
finished just under a minute faster than the HP system.
I didn't have any manuals or other documentation for the system, and so I
just had to trial-and-error my way through. I'm sure I didn't even begin
to scratch the surface of the capabilities of the system, but it was
nonetheless interesting to poke around in a system that was different than
our timeshare system. After I had tinkered around with it intermittently
for a few weeks, I ran out of things to try, and stopped playing with it,
partly out of worry about getting caught somehow.
A few months later, I thought about it again, and when I tried to login,
it appeared that the user had been deleted or the password changed.
Perhaps my tinkering had showed up as a larger than usual bill for
services, and that resulted in the change. Or, maybe it was policy to
change the user/password every so often. It was fun to explore while it
lasted.
The Teletype service guy would be much more careful when he was dialing up
the system and entering his userID/password, and would ask anyone nearby to
please turn away while he was connecting up to the system. I have no idea
if it was my tinkering that caused this change in behavior, but if I was
nearby, I would comply with his request, so I never got a chance to gather
up another userID/password.
Some years later, my father's business, which was a precision machine
shop, had an account on a GE Timeshare System that had the "APT" "part
programming language" that was used to describe a machined part. When the
APT program was "compiled", it would result in a punched paper tape that
would be fed into a Numerically-Controlled machining center to actually
create the part. It was very expensive to use, and though my Dad did let
me log in once and look around a little, I didn't want to rack up charges,
and stayed away from messing with it. The service was discontinued when
MASTERCAM came out for the IBM PC and they bought a couple of high-end PC
clones and licenses for MASTERCAM.
Anyway, enough old memories. If anyone out there had experience with GE
timeshared systems, or may know of existence of any distribution media or
source listings of the systems, or perhaps has memories of using them, I'd
love to read about it. If you think it might be of general interest to
the list, post it to the list, but you are certainly welcome to send it to
me directly at
moc-dot-enesneb+at+bkcir (backwards with special characters spelled out
to hopefully prevent it being snarfed up by 5p@mm3rZ).
Thanks for reading, and best to all!
Rick Bensene
https://oldcalculatormuseum.com
Beavercreek, Oregon USA