The limits of provable security
I've never quite understood the objections that people have to cryptographic schemes that are provably secure. If you have a proof of security then one of two things must be true: either the scheme is secure or there's a flaw in the proof. There's no other possible case.
All of the technologies that Voltage's products use have such proofs. The Boneh-Franklin IBE, the Boneh-Boyen IBE and our format-preserving encryption technology all have such proofs that have been published in peer-reviewed research journals. Because of this, I often don't understand the questions that I sometimes get about the security of our technologies.
Here's a situation that I've seen more than once. Someone asks about the security of our FPE technology, for example. We'll point them to the paper by cryptographers John Black and Phil Rogaway that has a proof for the security of the scheme. The next question is essentially, "But why should I believe that it's secure?" At that point, I'm never quite sure what to say next. If you have a proof in front of you, then either the proof is correct or there's an error in the proof. If this proof is of the security of a cryptographic scheme, then either the scheme is secure or there's an error in the proof. As mentioned above, there's no other possible alternative.
My recent experiences have led me to believe that there's a fundamental problem with this approach, and that's because many people really aren't comfortable with the idea of a proof. I've seen many cases recently where people essentially accept P and NOT P at the same time and don't seem bothered by doing this.
Every time I see this, I start thinking about the logical implications of it. After all, if you accept both P and NOT P then you can prove that absolutely anything is true. "If 2 = 3 then all cats are dogs" is absolutely true, for example. But because many people either don't understand this or don't believe this, I've come to the conclusion that proofs of security really aren't that useful. They may help specialists make sure that their work is correct, but to non-specialists they really don't say anything useful.