Computers and the language of bibliographic descriptions

摘要

The conventional methods of composing bibliographic records, developed over a long period for manual filing systems, are discussed from the point of view of what can be done in a computer system. The tendency has been to adapt the highly skilled and time-consuming process of composing formalised bibliographic records in bibliographic language by introducing a further degree of formalisation and structure in order to enable the computer to interpret the records in a way which simulates manual searching. It is shown that if one approaches the computer handling of bibliographic records from the point of view of what a computer is capable of doing, rather than adapting and simulating manual methods, it is possible to dispense with virtually all skilled preparation of formalised and highly structured records.A working system on this basis is briefly described. A data-base containing over 40,000 bibliographic references has been compiled by simply transcribing the descriptive data found on the title pages and elsewhere on the documents, using clerical personnel with minimal professional supervision. This data-base has been in regular use for 1 1/2 years as an on-line library catalogue for document retrieval, as well as a retrieval system for subject enquiries. The system is based on the Status-2 software developed by AERE, Harwell.Retrieval from natural language text is not new as a means of interrogating data-bases from the point of view of subject matter. The principle is here extended to cover the purely bibliographic elements as well, and to serve the purpose of library cataloguing. The wider implications of the use of natural language bibliographic descriptions, transcribed without modification, beyond the application to individual libraries, are discussed.