It's not a pejorative view but it does indicate that, in the opinion of
the Forum, the idea or the design is not ready to be in the standard.
We can find for each JOD section some who disagree with its move to the
JOD. However, unless the section is voted back into MPI-2, we need to
agree that it is not part of the standard.
Obviously individuals are free to pass around copies of any document and
pitch any idea, MPI-ish or not. We're debating the official actions of
the MPI Forum at SC'96. I feel it's important for the Forum to send a
clear signal that JOD is not MPI-2, and ideas in it do not have higher
priority or an easier glide-path into MPI-2 than any other idea.
Some JOD sections were hashed and re-hashed, that's true. But in the
end they were moved to JOD because they still had nasty unresolved
issues. Others were directly moved to JOD because they needed much more
rethinking or just didn't fit with the scope of the current MPI work.
JOD is a place holder for potentially nice ideas or incomplete designs
so they are not lost. Sections that try to move back from JOD into
MPI-2 will still have the same unresolved issues to deal with, the same
problems to address. So JOD is not a place for users to do "window
shopping" and select functions to force back into MPI-2.
For these reasons, I am in favour of not officially releasing JOD at
SC'96. Free free to personally take a copy, show it to your friends,
and urge them to flood the 1-800-2ASK-MPI number with requests for this
or that function. :-) But I'd like the MPI Forum position to be:
"Here are the 1.2 and 2.0 documents to be standardized, please review
them and comment on them. Thank you!"
--Raja
-=-
Raja Daoud Hewlett-Packard Co.
raja@rsn.hp.com Convex Division