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. xA The new system, then, should be designed with the aim of providing a literature-searching service for scientists. It can also provide a means for the experimental investigation of documentation systems, in the environment of a functioning information retrieval center. Once the goal of designing an operating documentation center is established, many design principles of small-scale systems must be abandoned. The collection being searched must be large enough to be of interest to someone. This will mean at least 50,000 documents and probably 250,000. The response must be adequately fast to satisfy the users. This will mean the abandonment of the present batch-processing arrangement in use at most computer centers. People who are faced with a one-day delay when using an information retrieval center are likely to avoid it, since it might then be more convenient to go to the library directly in order to perform a rudimentary search. Furthermore, the advantages gained by a batching of requests (the ability to perform many searches at once) are maximized at approximately one core fLLll of requests, or 100-1000 requests. This is a load considerably larger than can be expected to be needed initially in one day in an information system. One might then just as well plan for the processing of each request individually. Accumulating requests fora period of one hour, as a possible compromise, would probably not be sufficient to accunnilate enough requests to gain real efficiencies in processing, and probably would antagonize many users. The goal, then, must be a time-sharing mode of operation in which each request is processed individually and in which an effort ig made to provide an answer in a matter of seconds to a user who remains at his console during the search.