Back then, we had a 2-processor machine. It had to support multiple CICS
test regions along with the TSO users and batch. Online compiles were
not allowed because TSO was for short-running transactions with plenty
of think time. Tieing up your (possibly shared) terminal (and not doing
anything else while it ran) was a huge waste of precious resources
(machine and human). It was more cost-effective to do good desk checking
and compile in batch. With much cheaper (relatively) machines, letting
the computer do the debugging is the better choice now.
On 8/18/2021 12:17 PM, Tom Brennan wrote:
Me too, but in the early 1980's. I'd run the assembler from TSO READY
so I wouldn't have to wait for an initiator. My way of programming
was always like starting with a ball of clay generally like what I
wanted, then adding the details as I went along. That method means
lots and lots of compiles. Then one day my supervisor dropped by my
desk with a blue-bar listing titled, "Top 10 TSO CPU Users" and I
think I was on the top. Oops.
On 8/17/2021 11:35 PM, Mike Schwab wrote:
Well, in the early 1990s, my system had 1-2 hour delays on compiles.
So while waiting, I wrote a clist to do the same thing. Allocate,
error handling, and deallocate of a single file took about 30 lines,
and a few iterations of debugging. So, once I had one file allocate,
I went through all the files, executed the program, and deallocated,
and proceeded with the next two steps. Got it working and would go
get a new cup of coffee while it ran instead of having to wait.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
