For what it's worth, I am rapidly coming to the conclusion that external
standards are our friends in this situation. The problem with defining The One
True Canonical Representation (or even two of them) is that there will always
be some system(s) for which it is a bad fit. So then everybody will fight very
hard to make sure that their particular favorite environment isn't the one to
get screwed and progress will be very slow.
So I would propose that we punt on this whole issue by adopting as many
externally-define data representations as we can get our grubby little hands
on, and make all of them optional as quality-of-implementation features. The
MPI Forum can then concentrate on putting together the best possible mapping
between MPI datatypes and XDR, ASN.1, and the next N standards to come down the
pipe. Which IMHO is all that we can and should be reasonably expected to do
well. This problem is much bigger than MPI, after all...
-- Eric Salo Silicon Graphics Inc. "Do you know what the (415)933-2998 2011 N. Shoreline Blvd, 8U-802 last Xon said, just salo@sgi.com Mountain View, CA 94043-1389 before he died?"