I would suggest that a cryptography mailing list or newsgroup is a better place for this discussion -- it is not of IETF-wide interest. None the less... "David J. Aronson" <dja2001@att.net> writes: > How do cryptanalysis programs know when they've got it right? That's what the unicity distance metric is about -- it lets you judge the probability that you have the right key. I'd suggest reading Shannon on this. The whole reason you can't break a one time pad is that the unicity distance is infinite -- you can't tell that you have the right key even if you guessed it perfectly. > Now, suppose you salt the plaintext with rarer characters, so as to > flatten out the distribution. Techniques like this were first brought to bear hundreds of years ago. I'd suggest a read of "The Codebreakers" by Kahn if you are interested in the topic. Such mechanisms (including padding with nulls, homophonic substitutions, etc.) are long since superseded but at the time provided additional security to the ciphers and codes of the day -- they are now primarily of interest to historians of the field, although that is not to say that nothing like them is ever discussed any longer. See, for example, Rivest's Chaffing proposal of some years ago. -- Perry E. Metzger perry@piermont.com -- "Ask not what your country can force other people to do for you..."