摘要：Present day programs of computerized information retrieval overvalue the importance of retrieving "facts" without either attaching a scale of importance to the material with which they deal or ordering information in any way which corresponds to the order of human thought. The limitations of classification by subject heading become especially apparent when a body of information becomes, through new insight, pertinent to a new area of thought. That body of information thereby acquires new subject headings: thus one sees that the system of retrieval by subject heading can never serve to aid fundamental discovery. The dangers of the present approach lie in their devaluation of traditional methods. Critical reviews are devalued, personal knowledge of the literature is devalued, and a false impression is created that knowledge is the same thing as retrievable information. This diminishes respect for that sort of personal organization of knowledge which alone can serve creative insight.
