W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > May 2012

Re: Test 0212 versus XML EC14N

From: Ivan Herman <ivan@w3.org>
Date: Sat, 19 May 2012 09:31:33 +0200
Cc: Toby Inkster <tai@g5n.co.uk>, W3C RDFWA WG <public-rdfa-wg@w3.org>
Message-Id: <32E58E53-82E4-4BDD-997B-1561182E97CC@w3.org>
To: Gregg Kellogg <gregg@greggkellogg.net>
I think we excluded those tests because we knew that it is not possible reliably to automatically test implementations. There are uncertainties around the behaviour of SPARQL engines in handling or not equality correctly, let alone the fact that RDFa implementations are very often at the mercy of some underlying XML toolset that does or does not do the right thing. This is all due to the 2004 version of the XML Literal specification.

The RDF 1.1 standard addresses these issues (and, I must say, the experience we had with RDFa 1.0 and now with RDFa 1.1 has influenced the group). Hence, for the time being, it is better not to stir this up in our test suite.

Ivan

On May 19, 2012, at 24:35 , Gregg Kellogg wrote:

> Yes, I think you're right, that this is not correctly testing XML EC14N. Further more, there is no RDFa 1.1 test for it (I don't really exactly why 0011 was removed, but it wouldn't have generated an XMLLiteral in RDFa 1.1 anyway).
> 
> I'm reluctant to add such a test to XML and XHTML1 before REC, but we can fix the 0212 for RDFa 1.0 and add a new one that works with XHTML5 and SVG.
> 
> Suggestions?
> 
> Gregg
> 
> On May 14, 2012, at 12:08 PM, Gregg Kellogg wrote:
> 
>> On May 14, 2012, at 3:56 AM, Toby Inkster wrote:
>> 
>>> On Mon, 14 May 2012 12:52:29 +0200
>>> Ivan Herman <ivan@w3.org> wrote:
>>> 
>>>> There are significant changes under way in RDFa 1.1 on this.
>>>> Essentially, on the RDF lexical value side, the requirement of C14N
>>>> will be gone, any well formed XML fragment is fine (and the equality
>>>> of RDF Literals will be pushed down to the value space of the
>>>> literal). 
>> 
>> That's great, but we can't change RDFa 1.0, and I don't believe we'll change RDFa Core 1.1. Most likely, this would go as an additional rule in HTML+RDFa 1.1, and we could exclude the XMLLiteral case.
>> 
>>> Yay!
>>> 
>>> But my main point is about the test suite for RDFa 1.0, where test case
>>> 0212 contradicts the old 0011 and the RDFa 1.0 spec.
>> 
>> What do you think should be produced in this case?
>> 
>> The input is:
>> 
>> <?xml version="1.0" encoding="UTF-8"?>
>> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN" "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">
>> <html xmlns="http://www.w3.org/1999/xhtml" version="XHTML+RDFa 1.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
>> <head>
>> 	<title>Test 0212</title>
>> </head>
>> <body>
>> 	  <!-- In RDFa 1.0, if a literal contains XML elements, and no explicit datatype is set, the result is XMLLiteral -->
>> 	<div about="http://www.example.org/">
>>     <h2 property="dc:title">E = mc<sup>2</sup>: The Most Urgent Problem of Our Time</h2>
>> 	</div>
>> </body>
>> 
>> </html>
>> 
>> I believe that XML C14N causes the inherited namespaces to be demoted down to the <sup> element. Shouldn't the expected output be the nodeset with the addition of the default and dc namespaces?
>> 
>> "E = mc<sup xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">2</sup>: The Most Urgent Problem of Our Time"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .
>> 
>> What do you think should be produced in this case?
>> 
>> Gregg
>> 
>>> -- 
>>> Toby A Inkster
>>> <mailto:mail@tobyinkster.co.uk>
>>> <http://tobyinkster.co.uk>
>>> 
>>> 
>> 
>> 
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Saturday, 19 May 2012 07:28:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 19 May 2012 07:28:19 GMT