- From: <noah_mendelsohn@us.ibm.com>
- Date: Wed, 17 Mar 2004 19:34:44 -0500
- To: Dan Connolly <connolly@w3.org>
- Cc: www-tag@w3.org
Dan Connolly writes: > It occurred to me that in a way, RDF/XML syntax takes > the principle of independence of elements to an > extreme. Indeed, but I think it's important to realize that this approach is just what you say: an XML syntax for another (I.e. RDF) model. I agree that the RDF model has many desirable characteristics, including at least several that make it an interesting basis for negotiating interoperation between software written at different times and to different assumptions (I.e. for tackling the versioning problem.) I would agree that it potentially is a richer and more promising starting point than XML trees and XML schema. On the other hand, I remain unconvinced that XML tools are usually the right prism through which to view such a serialized RDF document. Consider a document containing the serialization of an RDF triple that expresses [Noah(known by his URI), isAuthorOf, Document(known by its URI)]. While you can use XPath, XSL and or XQuery on the XML serialization, it's not clear to me that this is the architecturally robust way to extract the author for the document. It seems that if the data is fundamentally RDF, you usually want a query mechanism that's aware of RDF and triples. Similarly, I don't think that XML schema (or WSDL) would be a first class way of enforcing the rule that every such document must have a statement of authorship. It seems to me that using XML tools on such RDF serializations is about like using Grep on XML documents; both are useful tricks at times, but Grep is not usually the right way to extract information from an XML document and XML tools are not RDF tools. As some of you have probably heard me state too often: it's exactly this tension between two interesting architectures, both of which we are promoting to our users at about the same time, that most concerns me about the coexistence between XML and RDF. Still, I agree that it's interesting to look at RDF in our discussions of XML versioning. Maybe there's something we can indeed apply. -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 --------------------------------------
Received on Wednesday, 17 March 2004 19:36:19 UTC