The situation is slightly different from the previous case where we considered
all possible uses of optional arguments. We have chosen to use optional
arguments in C++ for status, even though the values used in C could have
worked (though I agree with the decision the Forum took on the C++ bindings).
We should consider the MPI_STATUS case on its own merits, without worrying
about other candidates. I believe that a good case has been made that we are
ALREADY on the slipperly slope with "special case values" (MPI_BOTTOM,
MPI_STATUS_IGNORE, MPI_ERRCODES_IGNORE, ...).
Bill