winscape 2321

At that time, DEC's PDP-11 operating systems were all over the shop, with DOS, RSTS, RT-11 and RSX in perma-battle, but 11-M version 1, which was pure Cutler, must have been about 1974. I was working at DEC's CSS in Sydney. We got a visit from John GIlbert (one of the early DECnet developers with Stu Wecker, who I never met, but from whom I nicked a ton of excellent code) about then. He had a pre-production copy of 11-M. It was big news for us at the time, because we were in deep poo having promised that RSX11-D would run in 32K to an insurance company that had ordered lots of machines. It would hardly do more than boot in 48K. Memory was *very* expensive. 11-M version 1 would run in 8K Words - for small values of run.

IIRC the round robin scheduler arrived in version 3.2 or thereabouts. That would have been 1976 approximately.

IAS was laid out on top of 11-D. Its I-O was more complicated internally compared to M, but both always looked the same to user programs. 11-D had no notion of 'fork', but I-O completion ran at slightly elevated hardware priority, and had a much messier relationship with the exec databases. Notwithstanding that, drivers were loadable far earlier than getting that facility in 11-M, and the system could be sysgenned without buttembling the exec from source.

The main user visible difference was that you could type ahead at the command line and your characters would eventually fall into the right process's input buffer, just like a proper time-sharing sytem. 11-M's MCR never had that. 11-M+ got a sad version of DCL much later.

I got around that MCR restriction by running 4 VT-100's at once. I was usually listening to Mozart on headphones as I worked, and had to take a lot of "Rick Wakeman" jokes on the chin.

I remember IAS as being better at timesharing, but mostly for rich people. It was a Bentley compared to 11-M's Ford Escort. They possibly smartened up 11-D's scheduler more than a little.

I got a fair bit of work developing and butchering drivers for both after I left DEC in 1976. I preferred 11-M, mostly because of XDT. Freezing to rest in a computer room debugging 11-D drivers on the front panel switches of an 11-70 was not my idea of fun.

-- To de-mung my e-mail address:- fsnospam$elliott$$ PGP Fingerprint: 1A96 3CF7 637F 896B C810 E199 7E5C A9E4 8E59 E248