I haven't read a ton of academic research in general, but in trying to understand CRDTs and concurrency, gritzko's paper on "Causal Trees"[1] struck me as incre

I haven't read a ton of academic research in general, but in trying to understand CRDTs and concurrency, gritzko's paper on "Causal Trees"[1] struck me as incredibly smart and clear in its thinking. Many of the other CRDT papers I read (even influential ones) were flawed in a number of respects: blurred lines between design and implementation, blatant mistakes and typos, hasty and unconvincing conclusions, an overabundance of newly-minted terms and acronyms, dense proofs lacking any concrete examples, unintuitive divisions between operation history and state mutation. The Causal Trees paper is also dense and also invents a bunch of new vocabulary, but the logic is completely consistent (to the point of being unified under a single metaphor) and clearly explained every step of the way. The data format is also very clever, and the paper spends a good amount of time following the practical consequences of those design decisions, e.g. the ease of displaying inline changes, or of generating a particular revision of the document. Weirdly, the paper isn't much discussed alongside the usual contenders. (WOOT, Logoot, RGA, LSEQ, etc.) [1]: https://ai2-s2-pdfs.s3.amazonaws.com/6534/c371ef78979d7ed84b6dc19f4fd529caab43.pdf https://ai2-s2-pdfs.s3.amazonaws.com/6534/c371ef78979d7ed84b...