> The point was moot when no one else could run the program
> (it relied on Maclisp and PDP-10 computers running the "ITS"
> operating system, meaning that it could run only at MIT!)
Though in practice Macsyma was only run on the MIT-MC and MIT-ML
machines, there were actually many other machines which *could have* run
it.
Maclisp ran on TOPS-10 and TOPS-20, Digital's standard commercial
operating systems for the PDP-10, as well as on Multics (a from-scratch
implementation -- written in PL/I -- on completely different hardware).
There was never more than a handful of ITS systems in the world (4 at
MIT and a couple in Sweden if I remember correctly), but there were
thousands of PDP-10s running TOPS-10 and -20. Most of them were running
in commercial environments, and probably didn't have enough core memory
to run Macsyma anyway. I don't know if pre-commercial Macsyma ever ran
on these systems. Commercial Macsyma did run on TOPS-20, though again I
have no clue how many actual users there were.
There were about 100 Multics systems worldwide. Macsyma was run on
MIT's Multics system occasionally (when a user ran out of space on the
PDP-10), but I don't know if it was ever run elsewhere. Multics had the
advantage of a far larger virtual address space, but Multics Maclisp
used it a lot of it -- four words (144 bits) per cons cell.
Of course, all of these machines were on the way out. Minicomputers
like the Vax and workstations like the Sun were the next wave.
-s