PHOAKS is distinguished from other recommender systems by two major design principles: role specialization and reuse.
The PHOAKS system contains six months or more of recommendations and associated data for about 1,500 newsgroups.
PHOAKS searches messages for mentions of Web pages (URLs) and counts a mention as a recommendation if it passes a number of tests.
We obtained FAQs by tailoring the basic PHOAKS message-filtering architecture to identify Usenet messages that posted FAQs.
PHOAKs consists of a general architecture for filtering information from electronic messages and a set of techniques for generating and managing dynamic Web-based interfaces.
Finally, we are exploring a combination of keyword search and collaborative filtering we call "community-sorted search." The basic idea is to run a query using a normal keyword search engine, then filter the results through the databases maintained by PHOAKS. Results are clustered by the newsgroups that mention them (thus, disambiguating the meaning of the query), then ordered by the frequency of mention.
1 PHOAKS is available at http://www.phoaks.com/phoaks/.