@steedman87 Not yet. This article has been influenced by our work on replikativ though and we aim to get there again. I just think that CRDT's as an abstraction are too object oriented, we have a collaboration with Martin Kleppmann, who formalized them on top of Datalog https://speakerdeck.com/ept/data-structures-as-queries-expressing-crdts-using-datalog
@whilo practically speaking, am I correct in mapping what this formalization calls “ops” to datoms in a fact store like datahike? Or perhaps they can be mapped to transaction data/results.
You can use datoms both to represent ops or state, but here Martin stores ops, yes.
@csm301 Cool, I think so too 🙂.