Heraclitus is an open source suite of Java tools for Semantic Web adaptation. The Heraclitus framework proposes the adaptation of the Semantic Web, based on web usage data. Starting from raw access logs, the users’ online behaviour is modelled and adaptations are proposed for the evolution of the site topology and ontology. The site topology evolves through the insertion of shortcut links and highlighting of popular existing ones. The site ontology is semi-automatically built with the use of the SVM categorization algorithm and evolves through the insertion of new relations between its concepts. Heraclitus is licensed under the GNU LGPL license.
I ought to read their papers (which are online on the above site), but I’d be curious how Semantic Web got mixed with adaptative web sites? Possibly because of the word ontology. But what if you have a web site without a predefined ontology? And why do you need an ontology to build adaptative web sites? And why do we need adaptative web sites: can’t we afford smart web designers? (This last one was a cheap shot.)