MONO91 Operational Considerations chapter Mary Elizabeth Stevens National Bureau of Standards alone to guide him. 1/ However, if the full bibliographic citation, perhaps the abstract as well, is to be printed out by machine, the problems of limited character set are even more severe. This problem is today being solved, in some cases, by separate operations in- volving sorting and assembly of the full citations and abstracts of the items indexed, sepa- rately prepared, for photographic reproduction or typesetting. Hopefully, this partial solution will become obsolete as automatic type-composition eq\1ipment and computer-pre- pared typesetting techniques become more generally available. Operational considerations thus involve the costs, the availability, and the limitations of equipment now usable for machine-generated index production. Schultz and Schwartz report, as of October, 1962. "There are two major bottlenecks in automated index production caused by inadequate equipment development at the present state-of-the-art: "1. There is no way of using automatic input of the printed page or the indexer's notes; "2. There is insufficient flexibility in the forms of output available for a computer -produced index. Both of these areas are being worked on by equipment manufacturers, and an early solution has been promised." 2/ In general, operational considerations of this type do not affect the appraisal of auto- matic assignment indexing techrnques, because these have not yet been developed to the point of practical application on any realistic scale. Moreover, the difficulties of problem definition and basic understanding of language and meaning yet remaining to be resolved are such that radical new advances in computer technology, associative memories, char- acter readers and pattern recognition devices may completely alter the picture before practical systems are ready for operational tests. Thus, for example, it is claimed: "It appears desirable to begin experimentation with automatic indexing so that solu- tions will become known by the time character recognition equipment will have pas- sed the laboratory stage." 3/ Similarly, Doyle suggests that the "present rate of solution of the intellectual problems of IR is sufficiently slow that these advanced devices will be in common use long before IR will truly benefit from their presence", and he urges that researchers proceed as though such machines were already with us. 4/ 1/ Compare, for example3 Montgomery and Swanson, 1962 [421], p. 366: "This study suggests that indexing should be based on more than titles and that a bibliographic citation system should present to the requestor something more than titles"; See also, in addition to references cited, p. 61, footnote I, IBM "ACSI-matic auto- abstracting project. .. ", Vol 3, 1961 [290], p. 89: "The use of titles in document searching without any additional abstract seems to lead to a high number of errors, i.e., accepting documents which should be rejected, as not enough informa- tion is available to judge the pertinence of documents." 2/ Schultz and Schwartz, 19b2 [531], p. 432. 3/ Levery, 1963 [359], p. 235. 4/ Doyle, 1961 [169], p. 3. 172