Crypto

I got my boss reading Cryptonomicon. So far he much likes it, which is good, since it's such an incredible book. Too bad Evan is so stubborn.

Anyway, we got to talking about crypto and whether developing crypto would be a good business idea. (Him being all capitalistic and stuff...) We decided that in-house crypto is the only good crypto, since by trying to market its sophistication and durability, one would have to reveal the mathematical algorythm. By doing that, you open up the option of brute-force attacking the thing, which with a closed algorythm, would not be possible. At the same time, if one doesn't release the algorythm, there is no way for one to verify the effectiveness in the code.

A lose-lose situation from a business point of view. The only types that win here are the military types - since they can develop, test and use internally.

A few weeks back, my interest in cryptography piqued again, and I was thinking of designing a new algorythm for personal use only. It would be a spooky RSA like thing, with a rotating key with a 10 seconds lifespan. Also, the algo would use a unique juxtaposition technique, that's proven to increase randomness, and create a much more stable cipher.

The real question is, but many bits do I make the key? I'm thinking that 2048 bit would be fun :P

Anyway, off to watch game 7...