ORBIT(ID:4550/orb002)Query language by SDC Places Related languages
References: The NLM information retrieval system consists of a modification of the SDC ORBIT programs and is identified internally as ELHILL. Extract: Introduction NLM on-line experimentation began in late 1967 using the ORBIT programs of Systems Development Corporation (SDC). The MEDLINE prototype (AIM-TWX) became operational in June of 1970. Whole text search experimentation on a limited collection of toxicological material was begun in mid-1970 using the Mead Data Central system since it was specifically designed for whole text search and retrieval. The success of this latter experimentation led to a competitive procurement for a whole text information retrieval system in early 1971. Mead Corporation, Informatics, Inc., and Battelle Memorial Institute responded to the request for proposal, but surprisingly both SDC and Lockheed submitted a no-bid. Evaluation of the proposals led to a contract being let to Informatics, Inc., in late March of 1971. The system proposed by Informatics was the National Aeronautics and Space Administration (NASA) version of the STIMS/RECON package with systems enhancements to facilitate free text search. During the succeeding year, RECON was modified and the TOXLINE films were built. The TOXLINE service was publicly announced and demonstrated in April of 1972, and service instituted in October of that same year. A policy decision by the NLM in December of 1973 led to the transfer of TOXLINE to the NLM system. Service was offered from NLM on April 1, 1974 and discontinued through Informatics on April 25, 1974. The NLM information retrieval system consists of a modification of the SDC ORBIT programs and is identified internally as ELHILL. All comparisons drawn in the remainder of this presentation are based upon a relatively brief period of one and one-half months of operation at NLM and a substantially longer period of RECON operation. Extract: ELHILL structure The ELHILL inverted file structure is very simple and straightforward. All search elements from all fields are contained in one alphabetically arranged, heterogenous listing in a paired file structure consisting of an index file and a postings (linear file record pointers) file. The index file contains the term in a 39-character field, a count of the number of postings and either the posting when there is only one occurrence of the term or an address of the postings list for multiple occurrences of the term. All postings are four-character records. For textual terms one and only one posting is carried in the inverted file for multiple occurrences of a term in a single record. The six classes of search elements or fields available under ELHILL are: authors, language, journal coden, year of primary publication, text terms from both the title and abstract fields, and Chemical Abstracts Service (CAS) Registry numbers. |