Design of a document filing and retrieval system

This paper describes a scheme for secondary-key-based data retrieval. The scheme is designed for a document filing and retrieval system in which documents are filed under a large number of keys, but queries specify only a few keys. As this work is based on microcomputers with virtual memory, the primary objective was minimization of page faults, with processing speed and simplicity of code also of critical importance. Initial experiments with the scheme indicate good performance with respect to both processing time and paging.