Debbie
The reason the supplied ISTINCLM may have been discarded and your own
introduced *may* be due to an oversight by the VTAM developers when type 2.1
support was introduced - way back in the late '80s. It took VTAM development
a while to introduce the DYNMODTB start option, the mid '90s I think, and so
the only way to provide mode table entries for dynamically created type 2.1
node CDRSCs representing logical units performing the *secondary* role in
subarea session stages was to rely on the supplied mode table ISTINCLM.
Here's a snippet from my notes on Type 2.1 node support:
<quote>
A CDRSC for the secondary session partner *cannot* initially be created
dynamically unless the Session Management exit is active with the Adjacent
Link Station Selection function. In this case all the operands applying to
secondary operation are relevant - and missing! The Session Management exit
function can select one adjacent link station but all other operands must
take the default values, which, most importantly, means that the VTAM
default mode table, ISTINCLM, must be used. This is actually a severe
deficiency of the Session Management exit Adjacent Link Station Selection
function.
</quote>
With the introduction of APPN, the CDRSC for the secondary session partner
is created dynamically using APPN mechanisms and not the Session Management
exit Adjacent Link Station Selection function.
I'm sorry that was a bit technical; the merger of type 2.1 node SNA
architecture (which includes APPN) and subarea SNA architecture has its
messy aspects.
-
I expect the effect is consistent with its cause, it's just you haven't yet
identified what the cause is and so it's apparently inconsistent.
You can probably experiment on one PC's 3270 emulator product, say PCOMM, by
modifying the rows and columns specifications. Thus assumes you are using a
mode table entry which causes TSO to go and "read" the customised
presentation space dimensions. A test with 24 by 80 and another with 43 by
80 may suffice. On the other hand, in these cases TSO may always use "Erase
Write" for the default and "Erase Write Alternate" for the alternate and
always use the "three asterisks" when the change occurs. If you can arrange
to use a mode table entry which has X'02', rather than X'03', in the
penultimate byte of the PSERVIC operand, that will force the use only of
"Erase Write" and so there will be no need for a change and no need for the
"three asterisks".
You should be able to observe the use of "Erase Write" , "Erase Write
Alternate" and the "three asterisks" using NetView Session Monitor (NLDM)
primary trace. You will also be able to check the presentation services
field of the BIND - but you'll need to use the HEX option I think.
Chris Mason
----- Original Message -----
From: "Debbie Mitchell" <[EMAIL PROTECTED]>
Newsgroups: bit.listserv.ibm-main
To: <IBM-MAIN@BAMA.UA.EDU>
Sent: Monday, May 21, 2007 7:19 PM
Subject: Re: VTAM question (***)
On Mon, 21 May 2007 17:11:41 +0200, Chris Mason <[EMAIL PROTECTED]>
wrote:
I can't say I'd ever performed this sort of research but isn't the need to
regenerate - including reassembly - customized tables one of the sorts of
issues covered by the documentation a systems programmer goes through when
planning a new release?
As it turns out, the version of ISTINCLM in use here is not IBM's default.
It was customized long before my tenure here and not documented. I will
need to do some checking with other people here to see what they remember
about the reason it was customized as well as what changes were made.
Since
no one is really complaining (at least not very loudly) about it, it won't
be a priority. And, since posting my original question, I have noticed
that
the behavior is not consistent -- which leads me to believe that it is,
indeed, related to changing screen sizes.
Thank you all for your help.
Debbie Mitchell
Utica National Insurance Group
----------------------------------------------------------------------
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