Is the Dropbox Datastore conflict resolution API just a broken, naive version of operational transformations? https://www.dropbox.com/developers/blog/56/how-the-dropbox-datastore-api-handles-conflicts-part-two-resolving-collisions …
@smarx if you have a flat k/v space, why not e.g. Lamport clocks instead of an OT-alike?
-
-
@bascule Lamport clocks would just determine ordering, right? That wouldn't help to merge changes in the case of collisions. -
@smarx Lamport clocks would provide a partial ordering of events. "Siblings" would still arise -
@bascule None of our conflict resolution rules take into account event ordering. IMO, that's the right thing to do. -
@smarx from the server side. You're punting to the client -
@bascule I guess my answer to your original question is "no," it's not a broken or naive version of OT. :-) -
@smarx and I guess my response to your approach is "lots of complexity for half the wins you could get out of that approach" -
@bascule I'm still not clear on where there's unnecessary complexity. The protocol is remarkably simple (except lists, which do require OT). -
@smarx you're in the uncanny valley between simpler algorithms and ones which allow you to sync hierarchical data structures - 4 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.