INFORMATION RETRIEVAL: A PRAGMATIC APPROACH

Abstract

Much of the work on ‘information retrieval’ during the last ten years has been concerned with improving the ways in which information to be stored in catalogues and indexes can best be represented by terms suitable for use as ‘points of entry’ during the actual process of searching. Such work has been carried out from the aspects of both classification and alphabetical subject headings, and those who have been interested in the first aspect have derived much of their inspiration from the principles laid down by Ranganathan. Interest in more efficient methods for the construction of alphabetical subject headings received encouragement when Mortimer Taube produced his ‘uniterm’ system of co‐ordinate indexing, since he appeared to have removed the difficulties commonly associated with the formulation of compound subject terms, by abolishing their use altogether. Single (or unit) terms only were used, the searcher being required, at the time of search, to formulate his own compound terms by choosing two or more suitable ‘uniterms’. At the same time, Taube introduced as his method of referring to the relevant documents a system based on the results of earlier work with punched cards by Batten (ref. 1)—and possibly others working independently—in which the only information given to the searcher about a document was the ‘uniterm(s)’ and a file number—not even the title was given.