Metadata and Provenance

We represent metadata using RDF, a W3C standard for the semantic web. RDF statements are triples <subject, predicate, object> and a URI is used to identify each element of a triple. A set of RDF triples forms a labelled graph can be serialised as XML. RDF schemas and fully fledged OWL ontologies can be used to semantically type the RDF and perform reasoning - from simple taxonomic inferences to full logical reasoning.

Whole RDF graphs can be named, persisted, and then retrieved using suitable RDF query languages. For instance in MIAS-Grid, a named graph can be the metadata associated to a whole workflow run (process provenance) or the (RDF-ied) metadata associated with a DICOM image (data provenance). We have developed a generic process provenance ontology for this as well as domain specific ontologies such as a DICOM ontology based on the DICOM specification.

See Also

• TavernaProvenance
• Architecture