How can the predictive properties of Burrows-Wheeler transform be used to make a substitution cipher that foils frequency analysis?
Probably right after Move-To-Front (MTF). Arithmetic coding or partial squares for growth ratios less than 2. (For 2, use Poe's method of randomly choosing among many codes for "1", etc.) Order-0 coding is sufficient? (Is there alternation after BWT?)
No comments :
Post a Comment