This is an OPS-DIR review of IMAP REPLACE Extension (draft-ietf-extra-imap-replace-02) imo - this proposal causes negative ops issues - i.e. it fixes potentially bad ops situations where the state of a message repository is not determinable due to the failure of one of a chain of commands. so it reduces potential ops issues that said, it seems to me that section 3.4 could more clearly state that no changes are to be made to the repository unless all changes are made - i.e., 'all or nothing' - the information is there but it is a bit convoluted and the use of a clear simple statement would help