Fields of Information on Library of Congress Catalog Cards: Analysis of a Random Sample, 1950-1964

T HE Information Systems Office (ISO) of the Library of Congress has as its mission the development and implementation of the main automation program for the Library and the co-ordination of all LC automation efforts. One of the primary activities in this effort is a system-development study concentrating on the central bibliographic operations, that is, acquisitions, cataloging, reference, etc.' This study is now in its early stages, and it is too soon to predict the actual system that will evolve. As an adjunct to this study an analysis of the potential uses of and problems involved in the machine processing of cataloging data was begun. One aspect of the analysis was the design of a preliminary machinereadable catalog record. The results of this work are in a report issued by the ISO as its "Planning Memorandum Number 3."I2 The design emphasized retaining the traditional organization of cataloging data wherever possible in order to hold to a minimum any disruption in present methods while at the same time allowing the flexibility of manipulating the data for a variety of purposes. However, it was obvious that the machine record would have to contain information that is not now recorded on the catalog card but that the human infers from reading the cards. For example, the language in which the book is written is stated explicitly only when the text differs from the title as given on the title page; thus the user of the card makes an inference about the language of the work which the computer cannot do. For efficient machine processing those implicitly stated characteristics that are judged likely to be highly relevant in machine searching should be made explicit. In thinking of the uses of machinereadable cataloging data, such objectives as printing catalog cards, compiling book catalogs, user services (such as a subscription service of machinereadable cataloging data for other libraries), compiling special bibliographies and reports, a central store for on-line query, etc., were considered. The study of the cataloging data seemed to suggest that a suitable machine-readable record should be variable in length and contain both fixed and variable fields. A field is a unit of information that is explicitly identified and named. The name is derived from the major category of content contained therein; for example, title, edition statement, imprint, etc. A fixed field contains information that is al' B. E. Markuson, "A System Development Study for the Library of Congress Automation Program," Library Quarterly (July, 1966), 36:197-233.