EuroWeb 2002 Conference (EW)
17-18 December 2002
The Semantic Web is growing in both size and complexity. The number of RDF vocabularies and data sources are both increasing. This paper shows how large RDF databases can be accessed in smaller and more manageable chunks called RDF Objects. RDF Objects encapsulate compound data structures to give applications a more granular view of a database. Applications can control the size and structure of Objects by specifying the Object extraction algorithm and filtering the returned data by vocabulary. Links between Objects, analogous to hypertext links, allow applications to navigate across distributed data sources and leave link-trails for other applications to follow. Information can be aggregated from multiple data sources by recursively following Object links. Resources without URIs in different databases are cross-matched using DAML and OWL ontologies. An HTTP implementation of an RDF Object Serverhas been developed and used to develop tools for W3C Working Groups.