Discussion:
[Simh] MagicSix
Lars Brinkhoff
2018-11-09 09:09:44 UTC
Permalink
Hello,

Has there been any attempt to locate the operating system MagicSix for
Perkin-Elmer 3200 series (previously Interdata)?

Some highlights:

- Written at MIT Architecture Machine group (now Media Labs).
- Ran on a handful of 3230s.
- Similar to Multics.
- Virtual memory, segments.
- Chaosnet is only networking option.
- Primary programming language PL/1, and/or "pl/l"?
- Hosted Emacs clone TVMacs with the SINE language.
- Had a Lisp dialect, MagicSixLisp...
- ...which was used to write ASAS, animation software used for TRON.
- Used for the 1979 virtual reality experiment "Aspen Movie Map".
Phil Budne
2018-11-10 02:35:22 UTC
Permalink
Post by Lars Brinkhoff
Has there been any attempt to locate the operating system MagicSix
Larry Stewart wrote:

As far as bits go, I don't have any, although I do have some
hand-drawn schematics of the color display I built for the
Interdata model 85 and my old undergrad thesis (color printing).

Andy Lipman forwarded my query to Gregory Steve & Paul Mockapetris
with the prefix "Any ideas for this poor soul?"

Paul Mockapetris wrote:

I have a few shards of the Magic operating system somewhere, but
not nearly enough to make anything interesting run. I don't know
if there's any history or photos from the era. If anyone has any,
I'd love to get them.

My recollection of the history is that Nicholas forbade any OS
work. So naturally, while he went off for a summer in Patmos, we
built one.

Alternatively, since the Architecture machine was not famous or
funded enough at the time, the resources we had were a limited
amount of CPU time donated by the IBM Cambridge Scientific Center,
some Interdata minicomputers, and a hard drive.. So as I remember
it, Steve Gregory and I started some software, while Andy and
perhaps Mike did the disk controller and shared I/O bus. We
called the system Magic for Mockapetris and Gregory's Interactive
Computer, since we couldn't build an operating system

Magic 1.0 allowed basically a user per machine with a shared fail
system.

While ludicrous by today's standards, it was probably state of the
art at the time. We could easily have published in the computer
science journals of the time, if only we had known we were being
cool rather than practical.

I remember Jim Taggart had some software he was working on that he
would keep on multiple paper tapes of source code, then debug and
accumulate paper tapes of patches, and then go back and redo the
source and rebuild. The editing and compiling and linking meant a
new build from source would take a couple days or more. Once we
had Magic working, Jim could rebuild in 40 minutes or so, mostly
with his feet up as the system grinded away. He was our biggest
defender when Nicholas came back from Patmos. (I also seem to
remember a huge sigh of relief when we got the ability to back up
the Magic disk drive to mag tape, after I had been backing it up
illicitly to an IBM timesharing system via a terminal emulation a
scheme at 15 or so characters per second)

I believe this was the summer of 69 or maybe 70. I was off after
that,

Anyway, I remember that this stone age stuff was replaced by
bronze or iron age versions. I seem to recall Magic 6.0 with 32
bit architecture and a PL/1 compiler, but somebody else will have
to tell that story.
Lars Brinkhoff
2018-11-10 13:10:22 UTC
Permalink
Larry Stewart wrote: [no MagicSix software]
Paul Mockapetris wrote: [I have a few shards]
Thanks for checking.

That's discouraging. I also heard from Mike Kazar he doesn't have
anything.

Loading...