A relational interface for heterogeneous information sources

The World Wide Web today consists of a vast repository of diverse information distributed geographically across different locations. In the near future, this wealth of information will constitute the major bulk of information resources for digital library systems. To provide integrated access to web and other kinds of information sources, we believe that it is useful to first classify the space of digital library resources into collection and reference spaces. Searching can be done by first querying the indices or metadata in the reference space, followed by retrieving the documents or data from the collection space. In this paper, we present an integrated SQL query interface for web search engines, bibliographic and relational databases. Bibliographic databases, web search engines and certain structured databases are important sources of information within the reference space. By extending the relational data model and the SQL language with new predicates, our integrated approach permits a user to specify complex queries to any or all of the three types of information sources simultaneously. This complex query service in turn improves the effectiveness of searching within the reference space