A revolutionary method for precision web searching will be developed and tested. The new method combines knowledge bases of meta-data with a unified searching strategy and intelligent filtering and processing of results. It uses three innovative techniques, called knowledge links, search wizards, and distillation, to achieve superior searching precision over existing methods. A prototype of this method, called Knowledge Extractor (KnExt), will be tested on a combination of the entire web using multiple search engines and "hidden" web sites, the former IBM infoMarket, commercial databases and news-feeds (including Dialog, NewsEdge, Lexis-Nexis and Dow Jones), and private information stored in Excalibur's RetrievalWare, Lotus Notes, Documentum, Verity's Information Server, and Oracle databases. Experiments will be conducted to identify the most useful extensions to the knowledge base and the best natural language and artificial intelligence techniques to integrate into the distillation process to improve the searching precision of the prototype
