Key points are not available for this paper at this time.
It has been increasingly recognized that science information systems have need of natural language processing. F. W. Lancaster, author of the National Library of Medicine Study of the performance of the MEDLARS system, spoke of this at the 1971 annual conference of the ACM, in the panel "Can Present Methods for Library and Information Retrieval Service Survive?" He noted that "there is a definite trend away from large carefully controlled vocabularies and toward natural language processing, or at least machine-aided indexing," and quoted Klingbiel's remarks to the effect that "highly structured controlled vocabularies are obsolete for indexing and retrieval" and that "the natural language of scientific prose is fully adequate for these purposes."
Naomi Sager (Sat,) studied this question.