COMPARISON OF The PHILOSOPHY OF INDEXING TEXT WITH THAT OF INDEXING STRUCTURAL FORMULAE

At the Chemical Society Research Unit in Nottingham we are providing a service to some two hundred and sixty research chemists in which we search magnetic tapes produced by Chemical Abstracts Service of Columbus, Ohio (CAS), and which appear in a published form as Chemical Titles (CT), and Chemical‐Biological Activities (CBAC). In the course of this activity we are continuously presented with the problem of defining the interest of the user in a way which makes it possible to pull out from our information files those elements of information that are of most interest to him. In our particular system we do not rely on any pre‐indexing process, but rather input the data in a virtually unmodified form. When we are searching this information file we produce a search profile which defines the information of interest by means of a set of logically associated keywords or key phrases, which may be contained in the title, in the case of Chemical Titles, or in the digest material, in the case of Chemical‐Biological Activities, stored on the magnetic tape. Consequently my experience is most heavily directed towards that of searching free language information files, which have not had any great intellectual effort put into their input phase. Nevertheless I shall attempt in this paper to say something about the philosophy of indexing of text, and discuss whether substantial indexing is always necessary or even desirable, and to compare this process with that of indexing structural formulae, whose syntax and semantics are much less complicated than those of the English language.