FUTURE WORK

In building our prototype, the main categories under which extension document headings could be classified were hardwired into the code; for each of these categories, a table was created in the database. To enable our technique to work with any kind of document, we intend to remove any hardwired information and to allow for the definition of categories by the indexing user, i.e., by a user who knows enough about the domain and the documents. We will also extend our tool in order to enable it to automatically create any required DB tables and to dynamically generate the search interface. This will make our tool more generic and will enable its application in any domain.

Another area of future work that we intend to pursue is that of the agentification of the search component. By doing so, we will allow other agents within an agent-based framework to make use of it. For example, an expert system agent may use this service to link its conclusions with information available about these conclusions within the brochures .