We have now covered all the stages in the construction of the thesaurus using facet analysis as the starting point. It should be emphasized that, like any activity involving language, thesaurus construction is not an exact science.

Both the guidelines offered in the standards (which are largely concerned with the form of entry) and the methodology of facet analysis require the thesaurus compiler to exercise personal judgement in the selection and analysis of terms. Often the decisions to be made are not at all easy, because the subtlety and complexity of language does not always permit precise boundaries to be drawn. Don't let this deter you from attempting to construct a thesaurus. The application of method and logic will provide you with a good and efficient tool, even where decisions are necessarily subjective, and a little more experience of working with vocabularies will improve your confidence.

During the time that I was writing this book, the new British standard Structured Vocabularies for Information Retrieval (BS 8723) was in preparation, and the first two parts, Definitions, Symbols and Abbreviations, and the part relating to thesauri, were published shortly before Essential Thesaurus Construction went to press. In the USA, there is also a working party looking at the ANSI/NISO standard and considering many of the same aspects of the modern thesaurus not addressed by the current guidelines.

A workshop held at the end of 1999 set some criteria for the working party. The recommendations of the workshop1 included the following: