Re: [RDFa TC] Update from Toby A Inkster on 2008-05-27 (public-rdf-in-xhtml-tf@w3.org from May 2008)

From: Toby A Inkster <tai@g5n.co.uk>
Date: Tue, 27 May 2008 14:28:19 +0100
To: public-rdf-in-xhtml-tf@w3.org
Cc: ivan@w3.org, Manu Sporny <msporny@digitalbazaar.com>, michael.hausenblas@joanneum.at
Message-Id: <FCE16B54-7008-44E3-83D3-C94F36C62A12@g5n.co.uk>

> These b...dy xml literals again... And we always fall back to the same
> issue, namely that, although two xml literals may be identical in the
> xml infoset sense or, alternatively, in their canonical xml sense, the
> sparql implementations may not compare these properly.

Would a solution be to have the test case XML literals always result  
in RDF/XML fragments instead of XHTML or SVG as in the current test  
cases? Then the RDF/XML fragments can *themselves* be queried with  
SPARQL to verify conformance.

e.g.

	<title>Test Case 201</title>
	<!-- ... -->
	<div about="#test201"
	     property="[ex:testing]"
	     datatype="rdf:XMLLiteral">
	  <rdf:RDF>
	    <rdf:Description rdf:about="#embedded">
	      <dc:title>Foo</dc:title>
	    </rdf:Description>
	  </rdf:RDF>
	</div>

Then, instead of insisting that RDFa implementations produce this or  
that particular string as an XMLLiteral, just check that they do  
indeed generate an XMLLiteral; then feed that XMLLiteral itself into  
SPARQL and query whether #embedded has a dc:title of "Foo".

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Tuesday, 27 May 2008 13:29:28 UTC