Barbara Nitz wrote:
The lack of source code prevents me (as a customer) from
finding out fast what might have gone wrong and how to find a
bypass. And software support has gone down quite dramatically
in quality, in my opinion.
My worst experience along those lines came with MVT. When IBM
changed from simple console support to MCS, our system kept
crashing. The problem occurred when an operator issued a Mount
command with a seven character volume serial; due to the (new)
logic path the FREEMAIN for work space was issued twice. It took
only a few minutes to find the culprit in the source fiche,
whereas IBM's formal fix didn't arrive until months later.
I spent a lot of time working for service bureaus, and having
access to the source was invaluable in diagnosing customer
problems, as well as stopping some (e.g., TIME=1440 not
producing billing data - Seymour/Shmuel Metz found two places to
zap the Reader/Interpreter to 1439 by inspecting the fiche).
OTOH, ten years later applying maintenance to our Cobol compiler
caused 0C4s in the generated code, and IBM had a fix within two
days (error in base register assignment). Source wouldn't have
helped expeditiously, since the problem was too complex and none
of my systems group had any experience with the compiler internals.
Because of OCO, it's hard to tell whether IBM is doing a worse
job these days. Our early MVT system was IPLed every midnight to
compensate for storage leakage, etc., whereas the systems I've
used in the last twenty years tend to be available 24/7 except
for scheduled down time.
Gerhard Postpischil
Bradford, VT
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html