Fragments in RDF and HTML

In response to my outstanding action.

I have reviewed

Top-level summary: no problem, but a lot of dancing on pin-heads.

(it's only three or four paras)

Key statements:

First para - problem statement.

RDF uses an RDF URI Reference, which may include a fragment identifier, 
as a context free identifier for a resource. RFC 2396 [URI] states that 
the meaning of a fragment identifier depends on the MIME content-type of 
a document, i.e. is context dependent.

i.e. we would normally expect ...doc.html#frag to have a meaning 
depending on an HTML mime type, which is an anchor within a document, or 
some subsection of the document or something ...., but RDF doesn't take 
this point of view.

Second para - resolution.

a URI reference in an RDF graph is treated with respect to the MIME type 

i.e. wherever the fragment identifier originally came from, the RDF view 
forces a particular mime type.

Note that nothing here requires that an RDF application be able to 
retrieve any representation of resources identified by the URIs in an 
RDF graph.

So, in summary, RDF Concepts has the weasels words already prepared.

In these terms, an XHTML2 document and frag ID doc.xhtml2#frag when used 
within RDF/A is implicitly understood as if there was an RDF/XML 
representation retrievable with the mime type application/rdf+xml.
i.e. it is as if everyone who publishes an XHTML2 document has set their 
server up with the GRDDL transform and the appropriate mime type magic 
so that an RDF/XML representation of the fragment is retrievable; and it 
is that representation fragment rather than the HTML one that is being 
discussed. Not quite clear how you talk about the HTML document 
subsection as a piece of HTML rather than as an abstract resource, 
perhaps you can't really, or at least need an extended vocabulary to do 
so. The normal usage is that the doc.xhtml2#frag identifies a secondary 
resource which may be described (incompletely) in RDF, and have a 
representation in HTML, but is not, in itself, either its RDF 
description or its HTML representation.

What fun ...


Received on Tuesday, 27 September 2005 12:59:24 UTC