My friend Blake recently emailed me a link to a video tour of much gear brought together this past fall to celebrate the 50th anniversary of Data General's Nova minicomputer. I was a bit surprised not to have heard of it, not that I could have made it to Denver for two days of reveling in old systems.
The Novas and their successors in the Eclipse line had a good run. By the end of twenty years, though, a prescient eye could see the run coming to an end. About 1990, an article in Focus, the Data General Users Group magazine, mentioned a 386-based PC beating an MV/20000 minicomputer on a sorting benchmark. I would imagine that the PC cost at most $5000; the one reference I can find on-line for MV/20000 pricing says that they cost $200,000 and up when introduced in 1985. One could do a number of things with the minicomputer that one couldn't do with the PC, for example support a lot of word-processing users. But by the early 1990s organizations wanted to use the PCs for word processing. And by the middle of the 1990s, RISC-based systems running UNIX had largely supplanted the old minicomputers for such work as the PCs hadn't taken over.
I must say that I enjoyed working with the Novas and Eclipses. Given a day or so to brush up, I might be able again to write programs and scripts for them. Yet what then? Pretty much everything I did on the machines was for use, not play, and I doubt anyone is now doing the those tasks on them.
Yet here are to this day machines emulating the Eclipse instruction set to run old programs. How many, I can't say. Probably the people at Wild Hare Computer Systems have a good idea.
It was probably me, running ICOBOL benchmarks. And I made the 50 year celebration, along with my son, who got to see legends like Carl Alsing. Old strange times and anyone who says he'd rather have serial connections than CAT5 is a liar. :)
ReplyDeleteGood for you!
DeleteAgreed on CAT5: I learned a lot in troubleshooting RS-232 etc. connections, but I am used to modern speeds and would not care to go back.