W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > May 2001

Issue http://www.w3.org/2000/03/rdf-tracking/#rdf-ns-prefix-confusion

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Wed, 09 May 2001 17:45:07 +0100
To: RDFCore Working Group <w3c-rdfcore-wg@w3.org>
Message-ID: <30364.989426707@tatooine.ilrt.bris.ac.uk>
I was wondering about a related part of the above issue - defining a
meaning for non-prefixed attributes that aren't RDF concepts.

See attached test case (cloaked as HTML) [aside: can't we get this
fixed on W3C mail archiving software?]

Jan Grant already said something about related to this on www-rdf-interest

  Rather surprisingly, an unqualified attribute (in XML) doesn't inherit
  the NS of its element; nor is it implicitly qualified with the default
  NS. Instead, it lives in a global "unqualified" space. Yech.

  -- Jan Grant
  http://lists.w3.org/Archives/Public/www-rdf-interest/2001May/0107.html


XML NS[1] A.2 calls (non-normatively) the namespace of these
attributes the 'Per-Element-Type Partition'.  These are associated
with an element, not really global.  RDF does not have an
interpretation for dealing with these attributes, and they do not
match the current grammar.  I see two alternatives

  1) unprefixed attributes have no meaning in RDF and must not be
     used (excepting backwards-compatible accepting of rdf namespace
     attributes such as about, id etc.)

  2) the attributes are appended to the element URI and
     then follow the propertyAttr production

     In the above case we would get either of these statements:
     a)  file:example.rdf#node, http://example.org/bar#blah, "value"
     or
     b)  file:example.rdf#node, http://example.org/blah, "value"

     Depending on a sub-decision; is it a fragment or relative URI?


Checking some deployed code for what is done with example above:

repat   - dies: "unknown or out of context rdf attribute: blah"
rapier  - ignores it
SiRPAC  - generates a statement where the predicate is 'blah' -
          i.e. not an absolute URI, could be considered a relative one I
          suppose. [both SiRPACs]
VRP 1.4 - forms property URI http://example.org/blah i.e. takes 2b)


I prefer 1) since it seems the neatest way to remove the confusion
and is consistent with requiring namespace prefixes throughout.


Dave

[1] Namespaces in XML
http://www.w3.org/TR/REC-xml-names/



Received on Wednesday, 9 May 2001 12:45:19 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:35:46 EDT