From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 11/21/01
Let me try to answer the questions I think you are asking. Q: Your interpretations have several kinds of arcs. A: Not really, there are two kinds of information needed in an interpretation, one mapping vocabulary to entities (the bold dashed arcs in the interpretation picture) and one representing relationships between entities (the skinny non-dashed arcs). Q: What is the relationship between your interpretations and RDF interpretations? A: Our interpretations could be used as interpretations of RDF graphs. I think that there would be no difference whatsoever between our interpretations and the new model theory for RDF. Our interpretations can also be used to give meaning to XML documents that are not RDF/XML (because they violate the RDF striping requirement) and that thus cannot be turned into RDF graphs. Our interpretations do not, however, produce the same result for some RDF/XML documents because of what we consider to be broken parts of the RDF/XML serialization. On page 13 we give a formal relationship between our interpretations and the RDF model theory. Q: What other differences between XML and RDF do you capture? A: We allow for the relationships in the interpretation to be partially order, thus capturing both the ordered view of the world from XML and the unordered view from RDF. We capture the graph nature of RDF by annointing some attributes (and one element type) as logical (as opposed to non-logical) constants and giving these constants special meaning. We capture XML Schema primitive datatyping by having an XML Schema validation phase in our syntactic processing and using the result of that processing. We capture XQuery references by handing reference nodes. I hope that this answers your questions. peter From: Stefan Decker <stefan@db.stanford.edu> Subject: Re: followup from telecon (integration of XML and RDF) Date: Tue, 20 Nov 2001 22:38:16 -0800 > Peter, > > thanks a lot for the pointer to your WWW-paper. > Please allow me a question: > in figure 5 on page 12 you present an RDF graph and an interpretation graph. > The interpretation graph consists of nodes and arcs - as it seems, > several kinds of arcs. > Now it seems that I can go along, define the necessary vocabulary > (some arc names and node types) and represent arbitrary interpretation > graphs in RDF > (e.g., for arbitrary XML documents (with the cited restrictions)) > > Is this true? > If not, why not? > > Thanks and all the best, > > Stefan
This archive was generated by hypermail 2.1.4 : 04/02/02 EST