You ever work with WYLBUR?
Single address space, keeping users from crossing boundaries
(RACF, ACF2, Top Secret and WACF). Could edit a library with
RECFM=U. So one could keep source there if they wanted. Would, on
close compress the PDS to a single extent if it could.
Used very low level interfaces for allocation, such that SMS
would not even see the file get opened or closed. So I had to
finish fixing that so that in an SMS environment, that interface
could be turned off (in testing we found we could cause MVS to
have to be re-ipled), and then we used SVC99 for all allocations
after that (SVC99 takes a lot of resources as I recall).
Had its own scripting language, so applications were written to
run inside of Wylbur. With the SRB mode, we could read JES2 spool
directly (this was a problem, that I was going to fix when I got
to implementing SAF.... sigh.)
I have forgotten all the stuff that Wylbur did with stack
processing, and all so it could handle 250 simultaneous users in
one address space.
That was another thing I needed to fix. I needed to change Wylbur
Paging to use a larger number of pages to accommodate more users.
(yes, it did its own paging, and interestingly enough, CICS was
following along with what we did so that CICS/TS was doing what
we had just done with task management).
I absolutely loved working on Wylbur, best job I ever had after
Amdahl MDF.
Steve Thompson
On 9/7/2023 9:15 PM, Leonard D Woren wrote:
Bill Johnson wrote on 9/7/2023 1:05 PM:
We used to use ROSCOE at a small shop in the 80’s because it
used less resources. I hated it.
ROSCOE was one of a collection of TSO alternatives, which were
all junk. TONE, ACEP, Wylbur, maybe more that I don't
remember. They all had 1 two-pronged design goal: except for
Wylbur, a PITA in its own category, allow TSO-like online use
without the perceived overhead of TSO, and also, they would run
on systems other than MVS.
The reason the resource utilization of all of those was lower
than TSO is that it took longer for programmers to get their
work done, so the resource utilization was spread out over more
elapsed time, lowering the apparent resources used in a given
elapsed time period, but also lowering productivity. Something
beancounters generally don't factor because they don't
understand it. They liked the fact that a given set of
hardware could support 50 (choose your poison from above)
online users while TSO could support only 25.
Fortunately, we're way past hardware costing more than people.
/Leonard
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO
IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
