WWW Data Source Selection with SQLfi

World Wide Web is a great repository of information which requires declarative languages for querying it. In this context, an important problem to solve is the selection of the best data sources based in user preferences and application domain. For so doing it is necessary to manage a catalog with metadata describing data sources. Rigidity of traditional querying languages limits them to be used in this problem. Several extensions of SQL have been proposed in order to improve the expressive power of querying systems. One of them is SQLf with its extensions SQLf2 and SQLf3, which allow a great variety of flexible queries expressions involving user preferences, based on fuzzy sets. We have built a fuzzy querying system called SQLfi, which is an implementation for SQLf and its extensions. We propose to use SQLfi for building a catalog manager for data source selection in Web. It promises to be a more suitable solution to previous ones in both flexibility and performance