Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Did this 20 years ago in an FMMS. It's not worth it. It's better to have an arbiter source of truth that can guarantee ACID principles with transactions rather than introduce merge conflicts or lose transactions for lack of synchronization. The internet is almost everywhere, so use that rather than provide academic features that cause more headaches than they solve. SQL databases with transactions and row locks are invaluable inventions.

Also, if you want to collaborate, synergize, innovate, and revolutionize consider OTs. They're a known quantity. Handling merges of data is fraught with landmines.

You can't sell 2 of something to the same person offline and know if they wanted 1 or 2. Plus, giving an end user the ability to resolve merge conflicts is asking for theft and fraud.



Actually the theory to do this correctly didn't exist 20 years ago, so things truly are different now.


Sources and references help you make your argument shine!


Not a “proper” reference but according to Wikipedia (https://en.m.wikipedia.org/wiki/Conflict-free_replicated_dat...):

> The CRDT concept was formally defined in 2011 by Marc Shapiro, Nuno Preguiça, Carlos Baquero and Marek Zawirski.

Now, that’s not to say the gp wasn’t referring to using some of the CRDT concepts before they were defined collectively under that banner.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: