WSDL-defined XPointer fragment

WSDL defines a set of new XPointer fragments [1] for identifying
components in the abstract model of a WSDL document.  The primary
purpose of this new fragment syntax is to enable a URI to be constructed
identifying a component, specifically to allow one to make RDF
assertions about that component.  But it is also possible to use the
fragment during dereferencing.

We'd like the XML Core WG's input on an XPointer question.  For use with
the application/wsdl+xml media type (which we are defining), we would
like to use unqualified xpointer scheme names, such as:

  #interface(foo)

However, it is conceivable someone might want to deliver the WSDL
document as application/xml.

  #interface(foo)element(id-foo)

Is an unqualified scheme name acceptable in this case?  Or would it be
necessary to qualify the scheme?

  #xmlns(wsdl,http://wsdl...)wsdl:interface(foo)element(id-foo)

If the above is preferred for the application/xml media type, is there
any problem with having a different form for the application/wsdl+xml
media type?

  #interface(foo)element(id-foo)

[1] http://www.w3.org/TR/2004/WD-wsdl20-20040803/#wsdl-uri-references

Received on Wednesday, 29 September 2004 14:51:12 UTC