Hmm. another simpler theory to remove Shannon from the discussion. assume that the original assertion is correct - that for each plaintext p and each cyphertext c there exists only one key k that is valid to map encrypt(p,k)=c. In this case, for each possible cyphertext c, *every* possible plaintext p is a valid translation given a unique key k. for that reason, the uniary distance for encrypt() must be larger than one block - as it is self evidently not possible to map *any* c to a unique p without knowledge of the key. For that reason, Shannon cannot be applied to a single block of encrypt(), and can be safely ignored :)
--------------------------------------------------------------------- The Cryptography Mailing List Unsubscribe by sending "unsubscribe cryptography" to [EMAIL PROTECTED]