Discusses the problems of compressing a large textual database for storage on CD-ROM. A text-compression algorithm is presented, new algorithms for compression of indices are described, and the ARTFL (American and French Research on the Treasury of the French Language) database is used as an example. (14 references) (LRW)

Describes an approximate document ranking process that uses a compact array of in-memory, low-precision approximations for document length. Combined with another rule for reducing the memory required by partial similarity accumulators, the approximation heuristic allows the ranking of large document collections using less than one byte of memory…