Graphical database browsing

Abstract

The knowledge stored in the information networks of the near future will not resemble that in today's conventional database systems. Instead, systems will look more like electronic libraries, with millions of items in many different media and with widely varying formats and levels of detail. Using foreseeable searching mechanisms, the results of queries will often be very coarse, containing a large fraction of all the items in storage. To make up for this coarseness, users must be able to browse the database, discarding unneeded items rapidly—but the sheer size of the result sets will make textual presentation useless. As an alternative, we have investigated ways to arrange the data into information spaces that can be presented graphically. Users browse the space simply by moving their viewpoints within it, selecting interesting features of the “information landscape” for closer examination.