Machine Retrieval of Information

THEROYAL SOCIETY Scientific Information Conference in 1948 included a working party on mechanical indexing. This working party considered the various systems of mechanical selection which were brought to its attention: edge-notched cards; the Batten system; Hollerith and Power-Samas punched card systems; the Samian punched card system; the Rapid Selector; the Univac; Zatocoding; and the combination of punched cards and microphotography.l On the basis of the deliberations of the working party and its recommendations, the Conference concluded “that in the field of subject indexing and selection, designers of apparatus are well ahead of users in the facilities they offer, or plan to offer in the near future. The present need, therefore, is for experiments on a realistic scale using available appliances. . . .”2 There are two points which should be especially noted with reference to the considerations and conclusions of the Royal Society Conference on this matter of mechanical selection. In the first place, for almost ten years the list of available systems and appliances has remained constant. Anyone who today undertakes a survey to determine what is available in the field of mechanical selection will find almost the same possibilities considered by the Royal Society Conference in 1948. This seems a curious phenomenon in an age which prides itself on its rapid technological advances. But apparently this stagnation is not something which characterizes the machine designers. The second noteworthy point in the conclusion of the Conference is that the designers of appliances had advanced way beyond the willingness of the librarians, information officers, and documentalists who use manual systems to experiment with and utilize the appliances that are available. The Conference did not conclude that better machines were necessary but that those charged with maintenance of manual systems were lagging in their readiness to utilize