Introduction to the special issue on web-based information retrieval research

The tremendous growth of the World Wide Web and the increasing plethora of Web search engines and Web sites has broadened the scope of information retrieval (IR) research to include aspects of Web-based IR. Although the major focus of the IR community remains the annual TREC competition, a growing number of IR researchers are working with the tough problems inherent in designing real systems for real Web users. The IR problem is becoming more important to a large number of people who are discovering the reality of information overload when interacting with Web search engines. The issues of IR are also becoming more important to the scienti®c community as an acceptable topic for publication in prestigious scienti®c journals such as Science. A recent issue of Science includes large-scale Web studies, including Lawrence and Giles's (1998) groundbreaking study of Web search engine coverage and recency. Many of the classic issues challenging IR researchers from the 1950's remain with sharper dimension on the Web. Systems-oriented IR struggles with the development and testing of new algorithms, interfaces and systems. For many years the focus was on ``toy'' experimental laboratory IR techniques not scaled to commercial systems. With TREC the focus shifted to large-scale testing of IR systems using dichotomous relevance judgments by expert judges and static queries. Interactive TREC has emerged to look more closely at some interactive aspects of IR. The interactive IR community also struggles fairly independently of the IR systems folks, developing some models, but limited by a lack of large-scale users' studies and impact on systems design. With the Web, we now have a growing and large user base for research and experimentation. This issue begins to address the nature and problems of Web-based IR from the perspective of systems and user-based research.