Proposal: Strike section “Fragment Identifiers” from RDF Concepts

[[
PROPOSAL: Resolve ISSUE-37 and ISSUE-69 by deleting the section "Fragment Identifiers" from RDF Concepts
]]

Summary: This section only exists because of an apparent restriction in the old URI spec (RFC 2396) that was later removed by RFC 3986. Hence, the section no longer serves any purpose and can be deleted.



Detail:

RDF Concepts 2004 has a section “Fragment Identifiers”:
http://www.w3.org/TR/rdf-concepts/#section-fragID

This is an odd section. In general, the specification is very silent on the question *what* any given URI actually identifies. There is not a single scrap of information that would help us to find out what, say, <http://example.com/> could possibly identify. And that makes sense – the role of RDF Concepts is merely to define a data model. So why then is there a section that goes into excruciating detail on what URIs with fragIDs identify?

I think the answer is here, in RFC 3870, the media type registration for application/rdf+xml:

[[
In RDF, the thing identified by a URI with fragment identifier does
not necessarily bear any particular relationship to the thing
identified by the URI alone.  This differs from some readings of the
URI specification, so attention is recommended when creating new
RDF terms which use fragment identifiers.

More details on RDF's treatment of fragment identifiers can be found
in the section "Fragment Identifiers" of the RDF Concepts document.
]]
http://tools.ietf.org/html/rfc3870#section-3

However, both RDF Concepts 2004 and RFC 3870 are from the days when URIs where defined in RFC 2396. That document was obsoleted by RFC 3986, which is now also normatively referenced in RDF Concepts 1.1. And RFC 3986 clarifies this issue by explicitly allowing RDF's usage of fragIDs:

[[
The identified
secondary resource may be some portion or subset of the primary
resource, some view on representations of the primary resource, or
some other resource defined or described by those representations.
]]
http://tools.ietf.org/html/rfc3986#section-3.5

The same section also says everything one could possibly want to say about dealing with fragment identifiers when a resource has multiple representations (you should keep them consistent; if one representation doesn't use a particular fragID then that doesn't constrain the meaning of that fragID; existence of fragIDs doesn't imply that dereferencing has to take place).

So in my eyes, the section only existed because RFC 2396 could be misread in a particular way. This was later clarified and the section can now be removed. Hence my proposal at the top.

ISSUE-37 was about a late review of that section from Martin Duerst:
http://lists.w3.org/Archives/Public/www-rdf-comments/2003OctDec/0143.html

ISSUE-69 was about the problem that the section is closely tied to RDF/XML.

Both issues are largely moot if the section is removed, and could hence be closed

Both issues touch on one question that would still be unresolved: What happens when RDF is embedded into documents of other, non-RDF media types. But that's really out of scope for RDF Concepts – it just defines a data model, and is not concerned with the question what a URI identifies.

I pondered quite a while if the section can be updated so that it still says something useful, but the result would always be weird – why talk about <http://example.com/#foo> when the document is silent about <http://example.com/>? It should be sorted out between AWWW, Cool URIs for the Semantic Web, the RDF Primer and RDF Semantics.

Best,
Richard

Received on Wednesday, 9 November 2011 23:01:53 UTC