Mixing querying and navigation in MIX

Web-based information systems provide to their users the ability to interleave querying and browsing during their information discovery efforts. The MIX system provides an API called QDOM (Querible Document Object Model) that supports the interleaved querying and browsing of virtual XML views, specified in an XQuery-like language. QDOM is based on the DOM standard. It allows the client applications to navigate into the view using standard DOM navigation commands. Then the application can use any visited node as the root for a query that creates a new view. The query/navigation processing algorithms of MIX perform decontextualization, i.e., they translate a query that has been issued from within the context of other queries and navigations into efficient queries that are understood by the source outside of the context of previous operations. In addition, MIX provides a navigation-driven query evaluation model, where source data are retrieved only as needed by the subsequent navigations. This paper presents how MIX supports QDOM on views of relational databases.