User-oriented Index Design for Computer-based Documentation

Abstract

The traditional way of approaching indexing is to take a document’s special-subject terms as the starting point. The problem with this is that the resulting retrieval aid (the ‘back-of-book’/‘back-of-manual’ index, or the thesaurus/synonym aid for computerised retrieval) is not conducive to knowledge transfer (learning) when the user of the documentation is not a specialist in the field in question. Nowadays, many users find themselves in this situation, and are obliged to learn as they go along, from on-line or printed manuals. In order to make computer documentation more accessible to users, we have to take their knowledge needs and use of language into consideration. This is a relatively new approach in information retrieval. In the past it has tended to focus on improving precision and recall, but a more recent trend is towards improving the searching experience for users (Duffy et al., 1989; Ingwersen, 1994), with an emphasis on language (Blair, 1990) and knowledge (Ahmad & Griffin, 1993). Research on computer-assisted technical writing also now addresses the problem of indexing (O’Brian Holt & Williams, 1992). It may be more appropriate to talk of ‘indexes’ rather than ‘indexing’ when the emphasis is on the result of the procedure, and when the possibility of different indexes is envisaged to suit different users.