jurjen@cwi.nl (Jurjen N.E. Bos) (02/11/88)
In article 700 Adam J. Richter writes: > Here's a note for you cryptography wizards out there. >Assuming that there isn't some kind of bug in the program I used to >deduce this, I think I've found an interesting property of the >S-boxes. I have no idea how this might be useful and I have no idea >whether this property is a strength of DES, a weakness of DES, an >irrelevant artifact of DES, or a bug in my program. (Source code is >at the end.) > Consider the following question. For s-box S, input bit I, >and output bit O, how many input bit combinations that have input bit >I set result in output bit O being set, for s-box S? (Or, "how many >input-ouptput pairs for s-box S have both input bit I and output bit O >set?") He then tells us that this number is very regular indeed- exactly equal to the average (16) for the last two bits of any S-box. We have here in Amsterdam a real group of cryptographers, and spent some time at the problem. We did not find a reason why this would implicate the cryptographic qualities of DES (although it is likely) but we must admit that this is at least a remarkable coincedence. We did not find a bug in Adam's program. I strongly suggest that anyone who likes to struggle a little with bits plays with this, may someone will find something. PS about the simplification of the logical equations of DES: humbug. Those equations never will give a result that is more efficient that key search. Jurjen.