Indexing the World

The World Wide Web provides readers with a potentially effortless way to retrieve and view related materials from around the world. Although readers can reach many documents by stepping through explicit links, there is no convention to show what other related documents exist. Readers must know addresses of appropriate starting points for exploration. Indices (such as Jumpstation and ALIWEB) and Web roaming applications (such as the Mosaic-fish 'spider') are currently used to try and alleviate these problems. In their current forms these tools will only be useful while the Web is relatively small, and the user base tolerant. I propose that Uniform Resource Name resolution services provide an ideal location to serve indices. With reference to the work of Sheldon et al. [Sheldon94], I suggest that smaller indices linked to form a huge virtual index space provide a functional scalable method of indexing individual documents. The associative access provided by the system enables implicit links to be identified - essential in a dynamic hyperspace such as the Web.