Re: Public Defined

Rusty Lusk (lusk@mcs.anl.gov)
Tue, 07 Jan 1997 21:58:12 -0600

| My guess is, therefore, that a questionnaire won't help for a review of *THE*
| standard, though it could theoretically help to determine what people would
| like in *A* standard. It does seem a little late for a step like this, but
| maybe the current standard could be tailored to suit the results.

After MPI-1 seemed to be gaining some acceptance, there was indeed a
questionnaire, administered by the Ohio Supercomputer Center folks. It had
a number of parts, but one interesting part asked what people felt was lacking
in MPI. There were many responses, but most broke down into easily
recognizable areas: parallel I/O, dynamic process management, and one-sided
operations were the major ones, followed by C++ and Fortran 90 bindings, a
few requests for non-blocking collective operations, and some odds and ends
that are addressed generally by the external chapter. The results were
solicited and posted on comp.parallel.mpi, and also, I think, on
comp.parallel. There were on the order of a hundred responses. So I believe
that we are indeed addressing topics that people would like in *A* standard.

I confess that I am a little disappointed by the lack of response to the call
for comments, but I am not distressed by it. I speak often in public about
MPI, both 1 and 2, and the people I talk to seem happy about MPI-1, and happy
about the topics being addressed in MPI-2, without being familiar with the
current MPI-2 details. Rather, I believe that MPI-1 garnered sufficient
credibility that the MPI Forum is the beneficiary of considerable confidence
that it will do a reasonable job on the MPI-2 topics, which are indeed what
people want to see addressed. I hope we are worthy of such confidence, but I
have no doubt that we are working on the right topics.

I am not particularly concerned with the size of the document. There has been
a lot of work to do, and the Forum has done a lot of work. Just as MPI-1
exposed some issues that require some complexity for a complete treatment
(buffering, for example), so will MPI-2 expose some issues that require some
complexity for a complete treatment (1-sided, I/O, and external,
particularly). Maybe the presentation can be made more concise; maybe not.
Standards definitions are notorious for being bad tutorial documents; the
MPI-2 tutorial writings will no doubt emerge later, and will make it less
intimidating for "the public".

Rusty