WARNING: currently, there is absolutely NO security in demexp (as of versions 0.6/0.8)!

Why? Because before securing things, we need to have an absolute clear view of things to secure (i.e. requirements). Is is not currently possible without precise knowledge of the delegation subsystem.

We want to provide reasonable proofs that our users can trust demexp to keep there vote secret and to do proper voting algorithm.

A first attempt at defining security requirements.

• A very interesting paper on a low tech voting protocol without any use of cryptography! Unfortunatly unsuitable for Condorcet method. And several weaknesses are underlined in the paper. http://theory.lcs.mit.edu/~rivest/Rivest-TheThreeBallotVotingSystem.pdf And other weaknesses have been found: http://www.cs.princeton.edu/~appel/papers/DefeatingThreeBallot.pdf

• http://hal.archives-ouvertes.fr/hal-00142440 On privacy and anonymity in electronic and non electronic voting: the ballot-as-signature attack A very interesting paper where the ballot of complex voting schemes (like ranking in Condorcet voting) is used as a signature to identify the voter and thus void voter anonymity. Some counter-measures are proposed.

• Paper by Z. Gutterman, B. Pinkas and T. Reinman [PDF] on potential weaknesses in the Linux random number generator.

• Secure Programming for Linux and Unix HOWTO -- Creating Secure Software