Visual Querying for the Semantic Web

Principles of visXcerpt. The Semantic Web aims at enhancing data and service retrieval on the Web using meta-data and automated reasoning. Meta-data on the Semantic Web is heterogeneous. Several formalisms have been proposed. RDF, Topic Maps and OWL, e.g., and some of these formalisms have already a large number of syntactic variants. Like Web data, Web meta-data will be highly distributed. Thus, meta-data retrieval for Semantic Web applications will most likely call for query languages similar to those developed for the standard Web. This paper presents a demonstration of a visual query language for the Web and Semantic Web called visXcerpt. visXcerpt is based on three main principles. First, visXcerpt has been conceived for querying not only Web meta-data, but also all kind of Web data. The reason is that many Semantic Web applications will most likely refer to both, standard Web and Semantic Web data, i.e. to Web data and Web meta-data. Using a single query language well-suited for data of both kinds is preferable to using different languages for it reduces the programming effort and hence costs and it avoids mismatches resulting from interoperating languages. Second, visXcerpt is a query language capable of inference. The inferences visXcerpt can perform are limited to simple inference like needed in querying database views, in logic programming, and in usual forms of Semantic Web reasoning. Offering both, inference and querying, in a same language avoids e.g. the impedance mismatch, which is commonly arising when querying and inferencing are performed in different processes. Third, visXcerpt has been conceived as a mere Hypertext rendering of a textual query language. This approach to developing a visual language is fully new. It has several advantages. It results in a visual language tightly connected to a textual language, namely the textual language it is a rendering of. This tight connection makes it possible to use both, the visual and the textual language, in the development of applications. Last but not least, a visual query language conceived as an Hypertext application is especially accessible for Web and Semantic Web application developers. Further principles of visXcerpt are as follows. visXcerpt is rule-based. visXcerpt is referentially transparent and answer-closed. Answers to visXcerpt queries can be arbitrary XML data. visXcerpt uses (like the celebrated visual database query language QBE) patterns for binding variables in query expressions instead of path expressions – as do e.g. the Web query languages XQuery and XSLT. visXcerpt keeps queries and constructions separated.