Named activation groups vs *CALLER (was: H ActGrp('QILE') vs H ActGrp(*caller))

Synopsis: Outer (menu-level) programs use *NEW and all others use *CALLER

After reading such cogent remarks in favour of this plan, I'm a bit
concerned that my plan has been quite the opposite: All my ILE stuff has
gone in to named AG 'QILE'.  But I have a very mixed environment with mostly
OPM and very little new ILE.  The ILE stuff is quite segregated from the OPM
stuff in that it is all server style, where the interactive parts
communicate to a NEP running in batch via data queue.  In this limited
deployment scenario, AG(QILE) has served me well.

What I'm hearing is that I need to very carefully re-think that approach if
I intend to make mods to the existing OPM base.  The main issue is that I
wasn't going to convert an application en masse; rather, I'd tackle each
program as it needed maintenance.  This looks like a problem if I only
convert 1 program out of say 6 that comprise order entry.  If I make the one
converted program *CALLER, that'd put it in the DAG.  I'm not at all
convinced I want that.

It looks like I need to set up some test scenarios.  Thanks for the thought
provoking thread!

  --buck