They are involved with applying cutting edge research to difficult technological problems (see e.g. http://galois.com/project/proceed/ http://galois.com/project

They are involved with applying cutting edge research to difficult technological problems (see e.g. http://galois.com/project/proceed/ http://galois.com/project/proceed/ ). I don't think this paper is an example of that, but more of a way to issue an expert opinion - supported by a demo - on the risks of what is unfortunately not an uncommon practice. It might very well be obvious to everyone in HN why you don't send votes over any clear-text unsigned format over the internet, but unfortunately it is not obvious to the average voter or politician. Note that just encrypting the link is not really a good fix in this scenario, since that makes the server a single point of failure which attackers might try to subvert if the stakes are high enough. We have indeed solutions even for that (client-side signatures), but for really high stake elections we run into much bigger problems: the need to protect vote confidentiality from the polling authorities themselves, avoiding third-party verifiable proofs-of-vote because it allows vote buying/selling, allowing first-party verifiable proofs-of-vote since otherwise fraud is hard to detect, risk of fraud via compromised clients, etc.