Most protocols give you stereotyped headers, which are perfectly valid for known plaintext attacks. The rc4 cracks were done on the Netscape rc4(md5(key+salt) used in ssl. They were based on known plaintext in the HTTP headers. (Incidentally, we might want to test the key distribution & reporting mechanisms on a crack of vanilla rc4-40, or another SSL crack. Cracking des will not be cheap, and we should do some test runs first.) Adam The Deviant wrote: | > For instance if you had a DES encrypted gzipped file. The first 2 bytes | > plaintext will be Ox1f8b. You'd only have to try to fully decrypt | Buy the point is to prove that DES shouldn't be used, not that it CAN | be brute forced. A known-plaintext attack doesn't show that. We hafta | attack something we've never seen. (i.e. talk Netscape, or some other | company, into generating a DES'd message, and keeping the keys safe) -- "It is seldom that liberty of any kind is lost all at once." -Hume