Thread by @SarahJamieLewis

Critical cryptographic vulnerabilities aside, @Scytl keep saying "mixnet server" as if that is an actual thing with meaningful properties and not a major vulnerability in and of itself that annihilates the concept of voter privacy in these systems.

https://twitter.com/VTeagueAus/status/1201318998660407296

Back to the decryption proof response. The defence is now that "we would have detected if there was an attack" which is grand when the main adversary these attacks focus on is the people telling you that you must trust they didn't exploit the vulnerability in their own system.

Either:

1. The math is such that only honest components succeed with ridiculous probability

OR

2. There are certain components which must be trusted to not fail for a fair election to take place.

You don't get to sell a system based on (1) and then fall back to (2) on failure

There are so many things wrong with these systems that pointing out any one issue risks inspiring hope that these systems can be fixed.

They can't be, not while the focus is on things didn't go wrong, instead of why failures were allowed to progress so far.

If the public are happy with trusting closed source, vulnerable, uninspectable software wrapped in layers of obfuscation and ceremony to protect one of the most important processes society has then so be it, but we shouldn't water down what is being sold and defended.

The whole world of e-voting is so absurd that everytime I tumble back into the rabbithole I feel like Alice watching people to line up to play croquet with flamingos and hedgehogs watching vendors and election authorities bend themselves into arches.

There is a joke in there somewhere about believing as many as six impossible things before breakfast...but regarding the safety and security of evoting systems used in the real world to elect real politicians, and not that funny.