Half-day workshop on Monday, April 10, 1995
9.30 am - 1.00 pm
Room: 47/051
The amount of publicly available information resources on the Web is increasing rapidly. As this trend continues, finding these resources becomes more difficult. Several systems, notably Archie, Jumpstation, Lycos, WebCrawler, RBSE Index, and Harvest have attempted to solve this problem by building indices and allowing users to search them.
These Web-wide indexing solutions do not, in practice, share information with each other. In addition, the databases often fail to provide users with specific or complete answers to their queries.
This workshop has two goals: to find ways that index providers can share indexing and Web-structure information, and to explore ways to improve the query experience for users. Both goals stem from the need to address the increasing scale of the Internet: as the size of the problem increases, we need to be more efficient at building indices, and better at focusing on what the users are looking for. As we find solutions to these problems, it will enable us to build more efficient, consistent, and powerful indices of the World-Wide Web.
In addition to a general discussion of Web-wide indexing, the workshop will have two specific tasks:
Allowing indices to cooperate and exchange information can help solve several problems. First, it would allow the index builders to do a more efficient job of indexing. Indexing information for Europe might be collected in Europe, then transmitted in bulk to the United States. Or, we may find ways to build a distributed index, and avoid even the bulk transmission. Second, it would enable different retrieval engines to run against the same set of indexing information, providing better service for users and opportunities for research on different kinds of retrieval. Finally, it would server as an experimental tool for learning about decentralized indexing.
This workshop is not meant to set any standards for indexing or exchange of indexing information. However, if it serves as a starting point for the experimentation that will give us the experience necessary to propose standards in the future, should that be desirable.
The workshop environment is the ideal place to do this work; we will bring together people with experience building and running Web-wide indices. The ideal participant would be involved in building or operating an Internet information discovery system, or an expert in the field of database systems, distributed computing, expert systems, information retrieval, or library studies.