Various kinds of System reloads 2357

at least in the release 9.5-18 time-frame ... stage2 was more than a box (2000) but would fit in a tray (card trays held more than a box, but less than two boxes of cards, 3000?).

typical customer operation was to code up the stage1 sysgen deck ... maybe 60-100 cards. these were macro statements that then got "buttembled". the macros ... instead of generating real machine language ... were a bunch of "punch" statements ... which produced the stage2 cards.

stage1 tended to be relatively quick (modulo the actual punching of the cards) ... but stage2 could take an 8hr shift to run (or more). stage2 was a single "job" with possibly 100 "EXEC" (job) steps ... composed mostly of iehmove-iebcopy commands ... but also some linkedit and other misc. stuff.

starting with release 11, i would take the stage2 sysgen ... generate "JOB" cards for each of the steps (in part so i could run it in a production system ... rather than with the starter system). I also reodered various steps as well as iehmove-iebcopy statements ... as a means of careful placement-ordering of files and data on disk ... to improve disk arm motion.

I gave talks at share about both the work on optimal disk arm motion as well as doing sysgen in a production system. after i started also doing work on cp67 ... i talked about both the work on os-360, the work on cp67, as well as running os-360 under cp67 in a virtual machine.

references to talk i gave at aug. '68 share meeting in boston

--