Linearizability: A correctness condition for concurrent objects

This article, and several of the papers it cites, claim that (in the context of distributed systems) the property of "causal consistency" is an umbrella term without a precise definition. Which is simply false: causal consistency is well-defined and unambiguous.

The articles then define various, weaker, forms of what they (incorrectly) categorize as causal consistency, apparently to discredit Jepsen's analyses, and to support MongoDB's marketing claims.

Disappointing stuff.