I was having a hard time summarizing my thoughts from the Crypto conference. As always it was nice to see old friends and engage in discussions about the state of the art in crypto, but in retrospect I think the field of mathematical cryptography has dug itself into a hole and has a hard time seeing out. Lots of discussion about what paper got in, but not much excitement generated from the talks.
The invited talks were a pleasant part though. Gilles Brassard gave a nice historical account of the development of quantum cryptography. I always thought that it was a very nicely motivated topic, providing a nice alternative to the Shannon Information-theoretic model and the well worn complexity-theoretic model based on Turing machines. I was somewhat suspicious of the claims that it is economically viable, but customers generally end up resolving that one.
Adi Shamir gave a nice talk about algebraic attacks on crypto algorithms. His fundamental observation that crypto algorithms can be expressed as boolean polynomials provides a nice mathematical framework to work in, but it wasn’t clear to me from the presentation when it will be practical. I guess that’s part of the fun – trying to linearize things for analysis.
The rump session featured a nice talk on faith-based cryptography. It reminded me that many people in the field still throw around the term “provable security” in spite of the fact that the term is misleading. For most cryptographers their goal is to produce theorems rather than security, so they are completely satisfied if their output is “A implies B”, even if the majority of users of cryptography have little reason to believe in A. I guess you have to have faith.
In the end I concluded that I made the right decision to concentrate my creative energies outside of cryptography. Theorems are nice, but most humans have no use for them.