PODC/SPAA Invited Talk


Monday, June 29, 1998
9:30 - 10:45

How to Find It: Research Issues in Distributed Search
Udi Manber
University of Arizona

The World Wide Web has dramatically changed the way people share and acquire information, and its influence on business and education is expected to grow even more dramatically. Finding useful information is at the heart of things. Currently, people rely on central global search engines that attempt to capture as much information as possible. This approach has worked better than expected so far, but it is running into major problems of scale. Putting all the world's information into one flat database has its limits. I believe that the main trend in the future will be to distribute the search and to move as much of it as possible to the user's own machine, allowing more customization. A search in one's address book will always be preferable to a search in the whole universe's address book, no matter how powerful the latter search facilities become. Using a political analogy, this trend can be called the "democratization of search", because it moves from a monopoly of a few large entities providing all services, to a large number of entities from which users can pick, choose, combine, and customize. Giving users more control over how search is done and how they interact with the results is one of the major challenges.

This talk will discuss general research issues in this area, some recent work of the author, and many open problems.


This page maintained by Gil Neiger.
Last modified: April 27, 1998