DFTACTGRP(*NO) ACTGRP(*CALLER) was RE: More ILE 101

Like Colin, I've been doing it this way for quite some time.  I was under
the impression that it was a good compromise when dealing in an environment
that still used mostly OPM.  So I'm mostly doing calls from OPM to ILE.

After reading Bob's, Scott's, Joel's, and Jon's responses to the original
poster I'll accept that this is a bad idea.  But I don't quite understand
why....

Resources not being deallocated has been mentioned.  But are we just looking
at memory not being freed until the job is ended?  Files are still closed
correct?  I also thought RCLRSC had no effect, but Scott mentions that it
will cause variables to be re-initialized even if *INLR was *OFF.  Is there
any other effects?

When are resources deallocated when using a named activation group?

Biggest question...what do I use in place of DFTACTGRP(*NO) ACTGRP(*CALLER)
when most of my ILE programs are called from and may call OPM programs???

Should I use *NEW unless the program is called repeatedly?
Should I have all ILE simply use a single named activation group ('QILE')?

Thanks for the advice,
Charles