Multics Turns 40

Multics, one of my favorite operating systems recently turned 40. I first encountered it in ‘76 on the big machine at MIT. It was quite a revelation at the time. The other systems I used were either IBMs where you created cards on an 029 and handed your deck to the operator to queue up, or minis which used old style TTYs with paper tape readers. The Multics system, on the other hand, was like a visitor from the future. It served lots of different users while handling jobs at different priorities, and let you share things between users with different protection levels for different kinds of tasks.

My career has intersected with Multics people and Multics technology several times since then. When I was at Applicon, we did our VMS development in PL/1. We even wrote an interpreter for a variant of PL/1 called AGL. ArchMaq also used to use a strange version of PL/1 with a compiler that was written by a grad student. After Applicon, I went to Stellar. Bill Poduska was a Multician (he wrote the EPLBSA assembler), and number of the Multics ideas had been used in the Prime and Apollo systems. Of course, Multics really had an enormous influence on all of the computer industry because of something Bell Labs developed after they dropped out of the Multics project.

Multics is gone now, but you can actually get the source code if you’d like.

There are a number of articles around the web related to this anniversary (e.g. this interview with Corbato at CIO),  but you can also go to multicians.org and hear it all directly from the people who were there.