1) We didn't straw-vote on MPI_Accumulate_get, so this function might go. The removal would be very clean.

I'm with Rajeev on this one. I don't see the point. But, I think you knew that already ;-)

2) Should we allow MPI_NOOP in MPI_Accumulate (this does not make sense and is incorrect in my current proposal)

Agreed - it doesn't make sense to me either.

3) Should we allow MPI_REPLACE in MPI_Get_accumulate/MPI_Accumulate_get? (this would make sense and is allowed in the current proposal but we didn't talk about it in the group)

Yes, we should allow MPI_REPLACE - where else do we get a plain swap instead of a compare-and-swap? Keith