On Fri, 27 Jun 2014 13:51:51 +0000, Vernooij, CP (SPLXM) - KLM wrote:
>Yes!. In our group, we all have activated an 'initial macro' that sets for
>certain datasets some of the attributes we require, like STATS, AUTOSAVE OFF
>PROMPT (to avoid PF3's unintentionally saving modifications).
>
And I suppose it can even be project-sensitive, implied by "certain data sets".
>About what you call IBM chaos: if you look at the history of these features,
>it is explainable how things grew this way.
>
No, it's not explainable unless you tolerate irresponsibility. It could have
been
done right when the decision was made to do it at all. The function should have
been provided in Allocation so it would be available alike to ISPF, SVC99, TSO
ALLOCATE, BPXWDYN, and Batch JCL.
>Then we got problems with batch updating PDS's that were in use constantly by
>ISPF users, which was solved by batch updating the PDS with DISP=SHR, which
>corrupted a PDS once in every 1000 to 1000000 times.
>
Really!? Shocking! When I confronted the problem, with a naive caution I
resorted
to invoking ISPF LM services in batch IKJEFT01 to preserve integrity.
>Then PDSMAN intercepted disp=shr batch by adding the SPFEDIT enq and the
>linkage editor was doing likewise.
>
And even ISPF and linkage editor don't use the same ENQ. (I believe ISPF
uses both if RECFM=U.)
Alas, PDSMAN appears to be an ISV product, and can't be relied on to be
available in every customer shop. And we, as an ISV, may be compelled
to test twice, once for compatibility with PDSMAN, and once for integrity
absent PDSMAN.
>Now we have come to the chaos you describe and yes, someone with a supervising
>view could have stopped this trend and decided to do this at platform level,
>but this did not happen.
>
It's never too late.
>It is like a couple of guys developing a communication protocol for their
>computer connections that suited their needs, but now it is used as 'internet'
>and the total world economy depends on it, it should have been developed
>differently.
>
At least, unlike allocation/ENQ which is chaotic even within a single OS from a
single vendor, TCP/IP is uniform across all vendors (unless you regard SNA as
a viable competitor).
("couple of guys" seems to be devaluing DoD ARPA and NSF.)
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
