Foolproof query access to search engines
Bert Bakker and Johan ter Bekke
In order to support queries on the content of HTML documents and their interconnections, we propose to extend search engines with a visible data model of their index accessible via the Xplain language. Contrary to SQL this language only accepts queries strictly respecting the structure of the addressed data model. This language does not allow the specification of join operations. Consequently, malicious users are not enabled to execute extremely complex queries, possibly leading to a severe performance degradation of the attacked system. We show why in open systems as the WWW - with so many users - queries should not be specified in terms of SQL, but in terms of the Xplain language.
Keywords: semantic modeling, query language, internet, search engine.
Proceedings of the Third International Conference on Information Integration and Web-based Applications & Services (IIWAS 2001), Linz, Austria (2001), pp. 389 - 394.
Editors: W. Winiwarter, St. Bressan, I.K. Ibrahim
Publisher: Osterreichische Computer Gesellschaft, Wien - Austria, www.ocg.at
From the personal digital library:
Full paper in pdf-format [29 KB]