Formal methods have their place. I think we do need more waterfall-type design and documentation of cryptosystems (not massive pointless documentation, but rel

Formal methods have their place. I think we do need more waterfall-type design and documentation of cryptosystems (not massive pointless documentation, but relevant documentation and security/engineering tradeoffs design justifications.) Also, for more and better eyeballs to audit code, it needs to be simple and clear code. More thorough C unit/system testing, fuzzing and static/dynamic analysis steps as standard would help... I think a project/startup which could be like a TravisCI for the security and correctness of C code would work because there's just so much C code which needs better coverage.