Many thanks to
Maciej Ceglowski ("
lead troublemaker in the LSI revolution"); his swift response to my change requests for Search::ContextGraph means that I now have a working framework for retrieving - and persisting - document relatedness.
More on this very soon. In the meantime next-generation search wannabees should check out:
* The paper on
Spreading Activation Search by Maciej and his colleagues.
*
Search::ContextGraph on CPAN
* _Foundations of Statistical Natural Language Processing_, by Manning and Schultz (
Amazon UK link)
*
Introduction to Latent Semantic Indexing
* Tim Bray's
masterclass series on search engine development