RDF frag IDs and rfc3023bis

I have just been looking at:

http://www.ietf.org/internet-drafts/draft-murata-kohn-lilley-xml-00.txt
[[
5.  Fragment Identifiers

    Uniform Resource Identifiers (URIs) may contain fragement identifiers
    (see Section 3.5 of [RFC2396bis]).  Likewise, Internationalized
    Resource Identifiers (IRIs) [IRI] may contain fragement identifiers.

    A family of specifications define fragment identifiers for XML media
    types.  The fragment identifier syntax for application/xml is defined
    by two W3C Recommendations in this family, namely [XPointerFramework]
    and [XPointerElement].  Schemes other than the element scheme MUST
    NOT be specified as part of fragment identifiers for these media
    types.  In particular, the xpointer scheme MUST NOT be specified
    since it is still at the W3C working draft stage.

    When an XML-based MIME media type follows the naming convention
    '+xml', the fragment identifier syntax for this media type SHALL
    include the fragment identifier syntax for application/xml and
    application/xml-external-parsed-entity.  It MAY further allow other
    schemes such as the xmlns scheme and other schemes.

  ]]

When an RDF/XML document contains a reference to itself or another 
RDF/XML with a fragID, consisting of a bare name e.g. #foo then:
1) this refers to an rdf:ID="foo" (or equivalently an rdf:about="#foo")
2) this refers to the resource described by the document with that 
identifier, not the description itself, nor the XML infospace element
3) there are no XML IDs in either document

Thus the mimetype application/rdf+xml does not follow the SHALL above, 
since the "foo" fragID is specified under
http://www.w3.org/TR/xptr-framework/#shorthand
Either:
[[
An element information item may also be identified by an 
externally-determined ID value
]]
or
[[
If no element information item is identified by a shorthand pointer's 
NCName, the pointer is in error.
]]
is applicable, let's assume the former.
The fragID identifies an element information item, and not the resource 
described by the element information item.

This seems not dissimilar to the SVG case, where certain SVG fragIDs 
identify graphical objects in the image, and not the descriptions of 
those graphical objects in the SVG's infoset, as would be required under 
this I-D, and/or the xpointer framework.
http://www.w3.org/TR/SVG/linking.html#LinksIntoSVG
The wording there is careful to have #MyView denote the XML element, but 
it is treated identically to #svgView(viewBox(0,0,200,200)) which 
denotes a view (an abstract object)

FWIW my opinion is that URIrefs should have the same 
resource/representation duality as URIs... i.e. the specs may tell you 
how to get to a representation, but it is merely a representation of 
something else that remains hidden.
(Oh how I dislike metaphysics)

Jeremy

Received on Tuesday, 28 September 2004 16:58:04 UTC