Earlier this week, I gave a talk on LVars at RICON West, a distributed systems industry conference. It covered a lot of the same material as my talk at FHPC last month, but since my speaking slot at RICON was a little longer, I used the extra time to try to explain why a programming languages academic was speaking at a distributed systems industry conference, anyway. This post covers some of the same material in written form.
The short version: The mathematical framework of conflict-free replicated data types that distributed systems researchers have developed for reasoning about and enforcing the eventual consistency of distributed replicas is almost exactly the framework that we use in our own work on LVars for reasoning about and enforcing the determinism of parallel computations.
Update (December 13, 2013): A video of my RICON talk is now available. Thanks to everyone at RICON for being an amazing audience for this talk!