Object views through search views of web datasources

Abstract

Web datasources usually allow a restricted access (through CGI calls) and their output consists of generated HTML documents. Unfortunately, in many cases the data they provide happen to be available only on the Web. In this paper, we describe a system based on a Web wrapper combined with an object multidatabase system that enables the user to query Web datasources as well as other datasources with an OQL-like query language. The approach is based on two successive views of the Web datasource: a search view characterizing its access and an object view extending the restricted search view with respect the extraction capabilities of the wrapper. Both views are non-materialized to provide the user with up-to-date data. Our approach has been developed and demonstrated as part of the multidatabase system supporting queries via uniform Object Protocol Model (OPM) interfaces.