l.17: Group of processes -> group of MPI processes?
p.603
l. 1: Paragraph after binding starts incomplete without naming the binding again (old style?) ... the new text I am aware of always names the function again to give the sentence a proper subject. -> MPI_Win_attach attaches a local ...
l.22/23: with an MPI RMA call -> with an MPI RMA operation ?
p.605
l.28-36: Sentence first lists all attributes, then describes them in order "respectively" ... make the sentence hard to parse. -> One sentence per attribute?
p.606
l.6: Start of paragraph without subject
p.610
l.30: Start of paragraph without subject
p.612
l.31: Start of paragraph without subject
l.40: are incomplete, they must be completed. -> are incomplete, they must be completed separately.
p.616
l.21: Start of paragraph without subject
l.34: "A new predefined operator, MPI_REPLACE, is defined" -> This is not "new" in any sense today. The text should just state "a special predefined operator, MPI_REPLACE, is defined" (or similarly)
l.41: MPI_PUT is a special case of MPI_ACCUMULATE -> The difference between PUT and ACCUMULATE is state after this sentence, but this may stick ... maybe the "is" is too strong? -> MPI_PUT can be seen as an MPI_ACCUMULATE using the MPI_REPLACE operator. Note however, ...
Problem
References and Pull Requests
TBD