Better efficiency? What sort of efficiency are you seeking? I’ve built the core of a text OT system in C which can handle about 5M concurrent edits on a single machine. (Not counting network overhead.) I don’t think the upper bound on performance is the algorithms themselves.