Paul Gilmartin wrote:
On Thu, 21 Sep 2017 12:56:22 +0000, Dyck, Lionel B. (TRA) wrote:
Tone software in an advertisement from Enterprise Systems Media just claimed
that the CBT STEPLIB command and CA's TSOPLUS STEPLIB will not work with z/OS
2.3 in a move to sell their DYNA-STEP.
I have no problems with DYNA-STEP having used it in a prior life but I fail to
see how either of the other two solutions will fail with z/OS 2.3 as I don't
recall anything in the announcement or presentations that would imply changes
to the z/OS internals that would affect them.
Is IBM under any obligation to announce changes to the z/OS internals which do
not affect supported (GUPI) user interfaces?
<snip>
I'm not sure this has ever come up exactly this way before, so here is
an overview of what I think we do today in these cases:
If a code change does not affect the behavior of the system's externals
(including APIs), we feel free to change it at will without saying
anything externally. For example, on every new machine (it seems) we
need to keep HiperDispatch's topology view up to date and account for
other changes internal to the new processor. Except when we announced
HD to begin with I don't think we've mentioned any of these updates.
When we change the behavior of an external and you need to make a
corresponding change or use a parameter to keep the old behavior, we
think of that as a migration action, and put it in the book. (Marna is
the Keeper of the Migration Book. If you don't know Marna, go look up
the word "conscientious.") Sometimes the migration actions are "good"
migration actions, as for actions you will no longer need to take
because the system will do them for you.
When we remove an entire function, we generally announce it and try to
leave enough lead time for people to cope.
Outside public view, we have a program for telling other software
vendors about changes if they are members of PartnerWorld, which several
people have mentioned here in the past. When we suspect or know that
one or more vendors relies on something internal, we often notify them
of such a change. This is in addition to externals changes that are not
yet announced, but for which we believe they need some lead time. (When
we get that wrong, they are not usually shy about letting us know!)
The CBT tools are potentially affected by new releases, but there is
really no way to notify authors of upcoming changes in advance. Some of
those tools historically rely on internals, and they are most at risk.
Those written to APIs are less so but could be affected by a migration
action. (I'll note that I think it's really cool that many authors
update their tools and that others fix them up, all on a volunteer basis
and that Sam and Sam maintain the collection and the site.)
--
John Eells
IBM Poughkeepsie
[email protected]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
