ISR11 Design Consideration for Time Shared Automatic Documentation Centers chapter M. E. Lesk Harvard University Gerard Salton Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. X-2 to operational situations and make future plans. The following discussion is intended to be of general interest, although its specific proposals refer to the computing equipment (IBM SYSTEM 3&/67) to be available at Cornell and/or Harvard Universities in the near future. 2. Principles The first question that must be raised about any information retrieval system is that of purpose. For [OCRerr]Yhat users is the system to be designed, and what service is to be offered to them ? The SMART system has been an exclusively experimental system, whose user population is a small group of researchers at the Harvard Computation Center and whose function is to evaluate retrieval algorithms. It should become a system which provides practical information retrieval services for a user group of working researchers. This does not imply that no further work is to be done on small-scale experimental systems. There is much to be learned about retrieval procedures through detailed analysis of small examples[OCRerr]; new methods and systems will always be proposed that can be tried out economically on small subcollec- tions. However, work on large collections irnlst also begin. Operating systems are subject to various difficulties that must not be faced in experimental, small-scale systems. Experimental systems are used to investigate the details of retrieval methods in intensively analyzed document collections. This work remains valid, and in fact cannot be done in large-scale systems. However, the problems of user interaction, of time and presentation problems, of economics of use, of system maintenance,