On 12/9/2010 9:00 AM, Walt Farrell wrote:
While I'll agree they prologs are interesting to look at, I'm not sure we
view them as official documentation, and they certainly describe many things
that are not officially programming interfaces, with all that implies for
you (and potentially your program or system) if you try to use them.

Too often the books are incomplete or just plain wrong. Having time to file an
RCF to get them fixed is a luxury. Sometimes I have time; sometimes I don't. In
my experience, the macro prologues are the best place to find the accurate and
most up-to-date interface and usage details. Even the return/reason codes are
enumerated there and contain the latest information. The doc is often woefully
out of date.

Here's a recent true story: For a programming task started last month involving
Health Checker, the book was so inadequate that the programmer couldn't figure
out how to use the HZSQUERY interface to do what he knew should be possible. I
spent a good deal of time looking at it myself and finally asked, "Did you look
at the macro prolog?" Of course, what was needed was fully described there and
the programming task proceeded to the next phase. The new functionality is now
tested and working fine awaiting delivery to customers in the next release.

So anyone using the prologs -should- (in my opinion) be making sure that
what they're using is also described in the books or they're taking some risks.

If we weren't supposed to read the macro prologues, the z/OS programmers would
not create them in the first place or maintain them so painstakingly with each
change.

Though in this case I see that the prolog does indicate that USE2GTO32G is
not a programming interface, so one could get that info from a careful
prolog reading, too, as long as one read -all- of the prolog and not just
the keyword descriptions.

Exactly! IARV64 provides an on-point example of how the documentation in a
modern macro prolog is often _far_ more complete than the books.

The macros are carefully maintained by the highly-qualified z/OS programmers
writing the code in the U.S. or Europe or wherever they may be located whereas
the books are maintained by some non-programmer "ID" folks in a far-away land
called China.

The documentation the z/OS programmers provide is obviously intended to, and
absolutely succeeds in, helping us (the exploiters) be more self-sufficient. I'd
much rather read that information than waste everyone's time with days of
experimentation or a call/email to my friends in POK asking how an interface
actually works.

What would make us even more self-sufficient would be access to the source code
or compiler listing of the code behind the interface as is done in Linux--the
most popular emerging operating system in the world. Where listings still exist
for non-OCO z/OS modules, there are rarely any questions. In fact, experienced
programmers can often fully diagnose abends in IBM code as I did recently for
still-open PMR 06143,227,000 which resulted in still-open APAR OA34895. But, I
digress...//

--
Edward E Jaffe
Phoenix Software International, Inc
831 Parkview Drive North
El Segundo, CA 90245
310-338-0400 x318
[email protected]
http://www.phoenixsoftware.com/

Reply via email to