Benlog

Starting with some discussions at Crypto 2004, a few of us cryptographers discussed setting up new standards for evaluating voting systems.

We were specifically worried about the design-oriented, pass/fail nature of current standards. Instead, we want performance-oriented, multi-dimensional tests. Basically, instead of saying “machine X is good for voting because it uses DES encryption,” a fairly useless statement, we want a testing authority to say “machine X receives a 7.6 score on confidentiality and a 3.5 score on usability because 95% of users maintained voter secrecy and 40% of users were able to express their vote correctly.”

So we got together and wrote up a paper in this month’s Communications of the ACM. If you want to learn more, visit the web site and sign up to the newsletter.