VolodyA! V Anarhist wrote: > Problem with anything random is that the behaviour would then not be based on > the file. If you insert a file and reinsert it again, you expect to get the > same > blocks (and hopefully the same key all together).
It wouldn't matter, if the heuristic is good enough at guessing whether the file is already compressed or not - everything is random. Anyway you could easily solve this by telling the user if it was automatically compressed or not, and then they could manually use the same settings next time they insert it. > >> Also, I should think video/audio compression is somewhat more sophisticated >> than dictionary? JPEG uses discrete cosine transforms... >> >> X > > - Volodya >
