Left wing blogs

Carrol Cox cbcox at ilstu.edu
Fri Oct 4 16:17:36 PDT 2002


Kelley wrote:
>
> also, most of the programmers i know--the good ones--are creative either
> with art or music. in fact, as i recall there is a significant relationship
> between affinity for music and cryptography.
>

Not cryptography, cryptanalysis. Cryptographers encipher or encode texts. Cryptanalysts try to read the text someone else has encoded or enciphered.

When I was at NSA in the early '50s there was a general feeling that those with musical and/or mathematical training made the best cryptanalysts. But then I was the one who discovered that the Czech Border Guard was using isomorphic insteand of true one-time pad key, and I was neither. The mathematicians didn't think the frequency of key groups I had worked out were significant, so I just had to paw through several thousand messages until I found a couple enciphered with the "same" key. I doubt that there was much "artistic" about the process. You just had to see patterns and be willing to slog away for several months trying to match them up. More like ditch-digging than writing a symphony.

Carrol

P.S. An example of isomorphic key would be the 47759 and 04423 (which you could get, for exxample, by rewiring an electric typewriter so 4 gave you 0, 7 gave you 4, 5 gave you 2, and 9 gave you 3. One-time pad key is key which is never reused. It is in practice and theory unbreakable. The Czechs apparently had typists sit all day typing numbers in groups of 5 digits each. And if you do that all day you occasionally fall into a rhythm, as the one above (the typpist typing LRRLR (two fingered)and moving across to the right. That had enabled us to decipher scattered traffic, but not systematically. But that did give us a lot of key to play with and divide up into what we called key families, which we could run frequency counts on and so forth on our very primitive monster of a computer.



More information about the lbo-talk mailing list