Bill -- Do we plan to have any time allocated in the April meeting to MPI 2.2, or are we still in a holding pattern waiting for 2.1? If we do discuss 2.2 in April, I'd like a few minutes to discuss the whole "const" issues surrounding the C++ bindings. A few people asked me last meeting to rehash the issues, etc. In short: my position is that *all* predefined handles in the C++ bindings should be "const" (reversing some of the [incorrect, I believe] errata) and fixing a few C++ API bindings to compensate. MPI::BOTTOM is the only exception, but we have discussed that one to death already. If we push all the 2.2 discussions out to the June meeting, no problem -- I just wanted to queue up the "const" discussion for whenever the 2.2 discussions occur. -- Jeff Squyres Cisco Systems