|Approximating document descriptors: what to do when a catalog isn’t available|
Cunningham, S. J. (1997) Proc Electronic Library Visual Information Research Conference,Milton Keynes, UK, pp 125-131.
The New Zealand Computer Science Technical Reports collection provides a central index to over 32,000 working papers distributed in archives around the world. The collection is not formally cataloged, and cataloging information is available for only a minority of the documents. However, we can access and index the full text of the documents-not simply the title and abstract, as is common in bibliographic databases. We are investigating techniques for using this expanded keyword access to the full text so as to create "approximate" document descriptions, to allow the user to carry out searches similar to (although not as precise as) those supported by formally cataloged systems.