I am a little out of touch with current encryption technology but back in the day, I explored how DES worked. The S-boxes are something I've never seen since in an encryption algorithm and it didn't take long to find out that there is _no_ information about how they work! In practice they are a giant lookup table, a binary blob, if you will, buried inside the algorithm.

Cryptographers never managed to find a compromise based around the S-boxes, AFAIK. But it's hard to prove a negative. We just don't know if the NSA had a way in that no-one else ever stumbled across.

Fast forward to today.

The mathematics of encryption is now several levels beyond what I can intuitively pick up. It is a seriously bizarre field. I'm sure there are a lot of pure mathematicians gently bemused that this has actual real-world value. It means that programmers implementing them need the help of a very small group of people to provide and explain some of the magic numbers. Some of those people work for the NSA.

I bet some of those very smart people are regretting that now. Guilty by association and all that.

Wade.

But it's not small either. Not for the time.

But I stand corrected about the prominence and providence of S-boxes. I must have been looking at leading edge encryption right at the point that researchers were abandoning the idea. :-) However, my point stands that the design characteristics of DES's S-boxes only started appearing long after DES was on the decline.

I also note that DES was designed to be implemented in hardware. Software implementations really need a few CPU features that are not common for any sort of speed.

Wade.