RDFViewer - Searching Semantic Web based views
The RDFViewer system provides mechanisms to present different kinds of data as searchable views in terms of the semantic web knowledge representation languages RDF and RDF-Schema. A presented view can be queried using the RDF-oriented query languages SparQL and RDQL, as well as SQL. Currently such view definition capabilities are provided for relational databases through SWARD (Semantic Web Abridged Relational Databases) and for Topic Maps through SWATM (Semantic Web Abridged Topic Maps). By providing meta-data, queries, and results using basic semantic web representations such as RDF-Schema, the RDFViewer system provides scalable access to large repositories from advanced tools that need high-level semantic descriptions of integrated information.
For a given relational database the user defines with SWARD a view of the database in terms of the RDF data model. Using the RDF-model both relational meta-data (schema) and stored data can be queried. The RDF-view of a relational database is automatically generated from two simple mapping tables in SWARD that maps relational meta-data to an ontology. Similarly SWATM allows the presentation of the contents of a TopicMap database in terms of an RDF-Schema based ontology.
Using SWARD, we are developing a virtual repository system that enables queries to RDF views of large relational databases storing Government documents and life event data. The SWATM system presents RDF-Schema views of meta-data to describe life events in terms of the XTM format.
An RDF view of a data source is defined as a large disjunctive query. Optimization of queries to large disjunctive views is challenging and makes it critical to optimize not only data access time but also the time to perform the query optimization itself. We have developed novel query optimization techniques based on query rewrites and compile time evaluation of subexpressions that enables execution of real-world queries to RDF views of relational databases and Topic Maps.
There is a more detailed overview of the SWARD system here.© 2007 Uppsala Universitet, Department of Information Technology, Box 337, 751 05 Uppsala, Sweden | This page is maintained by Tore Risch