SP500215 Knowledge-Based Searching with TOPIC chapter J. Lehman C. Reid National Institute of Standards and Technology D. K. Harman Appendix A COMPANY AND PRODUCT SUMMARY Topic is a commercial off the shelf software product line available from Verity, Inc. Topic search technology is a commercial adaptafion of ideas extracted from the research of Tong, McCune et. al., in Rule-Based Information Retrieval, which was sponsored by the U.S. Intelligence Community. Topic supports cataloguing, indexing and retrospective search of fixed collections, automafic search of newly indexed documents according to (user) predefined search rules (profiles), and disseminafion/notification based upon satisfied search rules. Documents may be batched for indexing/profihng, or processed automatically as they arrive. The Verity, Inc. market presence in content-based text search/retrieval is described in the Delphi, Inc. 1992 Industry Summary. The Verity Topic product line is considered to have in excess of a ten percent share of the market in commercial-off-the-shelf content-based search/retrieval products for personal computer to minicomputer environments. Verity was founded in April 1988. The Topic product was first licensed and installed by the U.S. Air Force in June 1987. Verity currently has over 650 installations and some 30,000 users. Many thousands of persons have received training from Verity on the Topic products. Approximately one-third of Verity's installed base uses an event-driven or batch automatic-search-notification function. Many organizations use the routing mechanism for users who are unable to compose the (appropriate) queries, but require the expert's result quality. The Topic product line supports nearly twenty varieties of the UNIX operating environment, VMS, 0S2, DOS and MacIntosh. The product operates on data stored in the filesystem or in any SQL-based data base management system. The product as shipped supports over twenty formats of native data (markup languages), and provides the ability to insert local/third party markup language interpreters as required. A document in Topic is logical, and may be a file, subfile or any logical decomposition of a physical native document. The Topic end user (search) product is available in MSWindows, Presentation Manager, X-Windows-Motif, Macintosh, and character (keyboard/terminal) interface styles. There is a 40L-like command interpreter language for rapid applicafion development and remote command line interactive index/search. There is an Application Program Interface (C- library) to all Topic funcfions for embedded applications. 221