A PADRE in MUFTI (A Multi User Free Text retrieval Intermediary)

The Parallel Document Retrieval Engine (PADRE) has hitherto lacked the ability to support multiple time-sharing searchers, a deficiency which detracted from its cost effectiveness. Extensions, adjuncts and performance improvements are now proposed to permit multiple use and to ensure reasonable response times. This paper describes the proposed multi-user architecture, reports query-processing speed-ups, outlines a number of alternative types of user-interaction client including an interface to network browsers in common use. Implementation progress is reported and potential load handling capacity is discussed. KEYWORDS Text retrieval, information retrieval, parallel computing. 1 Introduction The Parallel Document Retrieval Engine (PADRE) has previously demonstrated that full text scanning methods supported by parallel hardware permit powerful query constructors and rapid response to changing document collections [3, 5, 6]. The addition of paralleldisk -resident inverted file indexes ..