There's a massive academic paper dedicated to Matrix's theoretical correctness (and the correctness of the Synapse implementation) going to be published in a few weeks. Wish it was already out, because it's spectacular :)

That said, nobody's done any Jepsen tests yet as far as we know, but we'd really welcome them!

Getting a Jepsen test done involves paying Aphyr for it, which I imagine you could do with this lovely new investment. Information here: https://jepsen.io/analyses

The first thing we earmarked the new investment for is doing a proper audit on the full stack - not just libolm (the e2ee ratchet implementation), but ideally an end-to-end audit of a typical best practice setup (e.g. E2EE-enabled registration/login/msg flows for Riot Web <-> Synapse <-> Riot Web), so we can then stamp a LTS label on it. We'll want to finish sorting teething problems on the new E2EE stuff though, as well as the scaling mess on Synapse first.

Sounds like Jepsen could certainly be part of that; we'll ping Aphyr when the time is right.

That is great news! Please post a link to a preprint when you can.

looking forward to that