Search

Computer searching of an information file offers significant advantages over manual searching when it can utilize strategies that are impractical for a human searcher to use. This paper examines several existing strategies, and then suggests that increased emphasis be placed on one that is relatively tolerant of "imperfect' indexing. Since "perfect" indexing does not exist, such a strategy would have considerable value. The general procedure consists of counting the number of search-matched index terms and then calculating a relevance factor for each item found, thus enabling the final search-output listing to be prepared in order of probable relevance. This technique is essentially the same as the strategy used successfully in such diverse tasks as identifying criminal suspects by their modus operandi, selecting suitable gifts for department store shoppers, and locating misspelled names in an airlines reservation file. [ABSTRACT FROM AUTHOR]

One of the applications of the Western Reserve University Searching Selector is the conducting of test searches on a file of encoded abstracts prepared for the American Society for Metals. This report presents typical results of pilot searches; also their evaluation as to pertinent interest and general usefulness has been made by the persons and organizations originating the information requests. The procedures used in making these evaluations are also presented. [ABSTRACT FROM AUTHOR]

The article provides information on automatic literature searching methodology. Traditionally, only the simplest tools have been used to locate and correlate information. Classification has relied on pigeonholes or equivalent compartmentalization. The mounting mass of recorded information has required the use of increasingly narrow subdivisions, necessitating that finer and finer distinctions be made in analyzing subject matter. Modern automatic equipment is able to scan and recognize index entries made in the form of holes on punched cards, magnetized spots on magnetic tape, transparent spots on photographic film, etc. Preliminary experiments with automatic equipment for information searching have shown that the key problem is the analysis of information for its essential features and the expression of these features in some medium. The uncontrolled or careless use of terminology is a particularly dangerous pitfall.

A letter from the editor is presented in regards to large data systems. The editor includes the paper "The Computer—New Partner in Investment Management,' by Dr. Arnold E. Amstutz, which describes work done on such a system where the effort of understanding how it would be utilized and organized spread over seven years. He then discusses the changes of data resources in the ten years prior to the article as well as their uses.

The set of different terms used to search a directory for information on a given topic varies considerably from one searcher to another. If the topic is represented by a written text, the characteristics of the text are among the variables which influence the formulation of search terms. The results of an experiment simulating a library catalog search indicated that a relation exists between the lexical content load of the text, used to define the search topic, and the number of query terms generated by a group of experimental subjects (college students). No connection was found between the content load of the text and the commonality of query term choices. Some of the properties of "popular," as opposed to "unique," query terms were examined. Only short terms enjoyed a high level of popularity. The repeated appearance of a term in the text favored the selection of that term by a large number of searchers. [ABSTRACT FROM AUTHOR]

The article presents the author's suggestion regarding the manner in which the use of a particular style of information search request could increase the efficiency with which the search results would serve the person making the request. To some extent conferences between searcher and search requester are used to frame the search definition in relation to the expected bulk of material it will yield. In most document systems, there are many headings that have some bearing on any typical search request. As the simplest example of a "figure-of-merit" search system, each of these headings could be scored as one point.