The RMA chapter introduces two new operators (MPI_REPLACE and MPI_NO_OP) as new:
A new predefined operator, MPI_REPLACE, is defined...
and
A new predefined operator MPI_NO_OP ...
Proposal
We should replace new because these operators are not new anymore and we should not assume that readers read the standard in chronological order (for which new would make some sense at least). A better term to use would be additional:
An additional predefined operator, MPI_REPLACE, is defined...
and
An additional predefined operator, MPI_NO_OP, is defined ...
Problem
The RMA chapter introduces two new operators (
MPI_REPLACE
andMPI_NO_OP
) as new:and
Proposal
We should replace
new
because these operators are not new anymore and we should not assume that readers read the standard in chronological order (for whichnew
would make some sense at least). A better term to use would beadditional
:and
Changes to the Text
Minor: replace
new
withadditional
.Impact on Implementations
None
Impact on Users
Less confusion, happier readers.
References and Pull Requests
TBD