Thanks for the detailed answer @metasoarous. I have replied on Discord as well, but let me also add here that Datalog itself is a very good implementation language for CRDTs, so the most flexible thing is that we collect some common merging (conflict resolution) strategies, query the datoms accordingly and add it to the db reflecting the merge. This is the idea is already sketched here https://github.com/replikativ/datahike/blob/336-native-image-cli/doc/cli.md#memory-model.

We have a new website btw. https://datahike.io/. Feedback is welcome!

That’s a beautiful website! And the code example (SQL vs Datalog) is brutally effective!

I really like the logo too. I suggest you put it on the README in the repository, a nice logo is known to attract users and contributors.