On 11/30/2005 11:24 AM, [EMAIL PROTECTED] wrote:
In that case, a multitude of customers would be delighted to see the date
and time symbolics moved from the class of "system symbols" to the same
class as &SYSUID, whatever you call the class. I understand (but don't
necessarily agree with) the concern over the ambiguity over system names.
But the difference between Reader time, Converter time, and Interpreter
time is different in character, and exists even now, even for STCs. If
an STC started just before midnight incorporates the date system symbol
in a data set name, that name may disagree with the timestamp stored with
the data set.
Most users don't care about such microscopic precision; IBM should relent
its intransigence and make at least the date and time symbols available
in batch JCL, with the stipulation that the values of such symbols reflect
the reader environment, not any other.
Unfortunately, since different systems in a sysplex (or sharing the
spool) could have very different times and dates, they could get very
unexpected results if we let those symbols be used in batch job JCL.
Suppose the job converted on a system where the date is today, and ran
on a system where the date is several days, or weeks, ago. (The normal
case is to have them very close, of course. And the more likely unusual
case is to have them simply a few hours off. But nothing requires that
the dates/times be anywhere close across systems, that I know of.)
Walt
----------------------------------------------------------------------
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
