* Oliver Bandel:

> I found an interesting paper there, about using updatable indexing
> ("Fast Incremental Indexing for Full-Text Information Retrieval"
> from Brown/Callen/Croft.) They talked about "inverted lists", and
> this together with other hints from this list may be a good starter.

If you need a full-text search capability on natural language
documents, there are various libraries you could reuse: Xapian, Lucene
(although this is one is writtein Java, IIRC), Estraier, OpenFTS, a
MySQL component, and probably many more.