don@allegra.UUCP (D. Mitchell) (11/24/84)
Letter correlations can be used to generate pseudo text. I have a simple program that does that using a markov process initialized by the Webster's dictionary: unfultie raderalize clobbler non aindtion impropyre hemic ramery hogimoingly igly alostyton harmatitela ereaeodeita rang incon al ovelegagets ingutradner ous ovansonnon hety eloetedam noneenobing ativindite hypness ene man obergy entrai tose crophoole res implancy iconitect ic omeny unal elia late echess over aited exanuraferrectous It is much better to base the model of text on words instead of letters tho. The following is from a markov process initialized with some Chinese philosophical texts: The tao will never pass away. The tao is absent from the very beginning until now its name has never been forgotten. Thus, a good cook changes his knife once a year because he does not collect precious things lead one astray. Therefore followers of tao. That is why a victory must be observed like a little child. Heaven and earth. This is not powerful, but they are in the silence of the needs of others.