Assigning Semantic Meanings to XML

The Semantic Web Project of W3C is developed at the purpose of extending the current web to make the information automatically processable by software programs. RDF/RDFS, DAML+OIL and the recent OWL are several specifications proposed along this way. However, the proliferation of XML on the current web makes it a tempting target if we can make use of these existing resources. XML is good at marking up the structural relationships, but it has limited capability in presenting the semantics contained in the data. In this paper, we describe our approach of extracting the propositional contents from XML documents and representing them in the ontology formalism. For any XML document, we first build up a surface ontology directly from its DTD file and encode the data in this ontology. We then represent the dataset in some more standard domain ontology via ontology translation. The Meaning Definition Axioms bridging over the surface ontology and the deep standard ontology have the function of assigning meanings to XML tags. The reverse direction which serializes a dataset of the standard ontology to a specific XML language works in the similar way. Combined with our former work, the whole system is capable of translating between the structural XML and the semantical RDF/DAML/OWL. It also can translate between heterogeneous XML documents about identical or similar domains of discourse.