Caveat: I've had about 2 secs to skim the proposal. Looks interesting. JSON Patch was new to me, and potentially very useful for work we're doing right now in

Caveat: I've had about 2 secs to skim the proposal. Looks interesting. JSON Patch was new to me, and potentially very useful for work we're doing right now in Myna. However there seems to be the assumption there is one client and one server. This isn't going to cut it if you want to do the fast or the reliable, in which case you want multiple data centre deployment, or if the client has device(s) that they might operate offline. In this case you have eventually consistent data, so you need some way to merge it, and this normally means you need some versioning scheme. I'm most of the way through this paper which seems to offer the best solution to CRDTs (a way of handing eventually consistent data) I've read so far: http://arxiv.org/abs/1307.3207 http://arxiv.org/abs/1307.3207 Monoids for ever!