> The mathematical proof might be 100% solid and hacker proof. The implementation will probably never be. I agree, I see a few ways that it could go wrong: -

> The mathematical proof might be 100% solid and hacker proof. The implementation will probably never be. I agree, I see a few ways that it could go wrong: - Code is proven correct against the specification, but specification is wrong/buggy - source code is correct/secure, but compiler is overzealous and botch the securities guarantees.