When they have specific information needs, users commonly express these in the form of a query or search statement and much research has concentrated on this area. Users may formulate queries in several ways, e.g. by:

  • Using natural language
  • Entering one or more search terms (‘keywords’)
  • Completing a search template
  • Selecting search terms from menus offered
  • Formulating logical statements (using Boolean logical operators to assemble queries)

The first two options share some common characteristics in that neither requires familiarity with specific database terminology. Although this can appear to make them more friendly users may be misled into believing the system ‘understood’ their search request. Despite this, simple keyword searching has become a de facto standard for a generation who use Google as their premier information resource. It should therefore be offered as a search option by any NBA targeting the general user population. However, since keyword searching typically results in low precision and large result sets, some relevance ranking mechanism should be implemented to improve.