The results of the 'wordlist' and 'semantic' backends, see RFE 1695128, and RFE 1695126, can be used in lots of interesting ways. This RFE is a reminder of some of them, and the processing required.
- spellchecking the words
- word statistics
-> find keywords for a document, set of documents
(Need algorithm references which do stuff like this
automatically - Bayes usable ?)