rexx or other macro processor on zos
 |
Paul Gilmartin
this was standard operation for "sysgen". you would punch up 40-100 cards which specified a lot of parameters for the system configuration (specified as buttembler macros). this was "stage-1" sysgen.
some more, this time from oct89, also mentions NSWC (naval surface warfare center) and NOSC (naval ocean systems center): An intense effort is centered on the evaluation of...
based on parameters and a punch of other stuff, the buttembler macros would "PUNCH" a couple thousand cards (frequently around a box). This became "stage-2" sysgen. This was job card, a whole series of job step EXEC cards and all the necessary control & specification statements for each EXEC step. it wasn't that it was a pbutting data ... it was a whole new job stream that had been generated automagically by the buttembly of the stage-1 sysgen deck.
this was done in dedicated scheduled time (typically at least a shift or two), booting-ipl a special "starter" os-360 (PCP) system that handled first the stage-1 sysgen followed by the stage-2 sysgen deck. A single stage-2 sysgen deck (couple thousand cards) would be a single job with several score job-EXEC steps.
as i've mentioned before, when I was an undergraduate in the late 60s ... starting with mft sysgen for os-360 release 11, I manipulated stuff so that I could do stage-1 sysgen in standard production job stream, get the resulting output for stage-2 sysgen and completely re-organize it. I worked it so I could submit it to hasp-mft production system with most EXEC steps organized as individual job steps ... but carefully controlling their sequence of end. The sequence of EXEC steps (and things like the order of move-copy statements and other stuff ... within EXEC steps) were organized to carefully control the physical placement of datasets and PDS members for optimal seek ordering on the newly generated system disks.
Anne & Lynn Wheeler My one comment: I come from a deep real-time background. If someone starts tabulating interrupt latencies or bandwidths, that's...
recent post commenting that they finally offered being able to explicitly place VTOC in release 15-16 ... something that I had been asking for since I started careful disk layout with release 11 system generation.
for other topic drift, several collected posts about early use of rex (long before it was renamed rexx and released to customers)