On Dec 3, 2003, at 7:48 AM, Henry S. Thompson wrote: > <?xml version="1.0"?> > <root xmlns:p="http://www.example.org/"> > <p:book> > <p:title>The Mill on the Floss</p:title> > <p:author> > <p:person p:title="Ms" p:given="George" p:family="Eliot"/> > </p:author> > </p:book> > </root> > > Now, just what Web Resource does the QName 'p:title' identify? None, because it is not a URI and there's not a canonical way of mapping it to a URI. > It seems clear to me there isn't one. The Namespaces REC provides a > means for defining _names_. It does _not_ place any requirements that > those names be univocal within a given context. This seems perfectly > reasonable to me. It follows, however, that the kind of names the > Namespaces REC defines cannot usefully be assumed to be in one-to-one > correspondence with Web Resources, even against the background context > supplied by a XML document type, however defined. We agree; XML+Namespaces makes resources out of vocabularies, but not individual elements and attributes. Cheers, Tim Bray http://www.tbray.org/ongoing/Received on Wednesday, 3 December 2003 13:57:49 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 12 September 2008 07:02:00 GMT