Report on the INEX 2004 interactive track Anastasios Tombros, Saadia Malik, and Birger Larsen   View Paper (PDF)   Return to June 2005, Volume 39, Issue 1 Abstract As scientific data repositories, digital libraries and publishers increasingly use the eXtensible Markup Language (XML) for publication and storage interest has arisen in exploiting this formatting for retrieval purposes. XML is attractive because it defines the logical structure of the documents and has the potential to assist IR systems in providing more appropriate results to users, i.e., to return relevant document components (i.e. XML elements) rather than whole documents. In addition, the XML tags often have specific semantics that may be exploited purposefully in IR. ©2006 Association for Computing Machinery