Re: Review

On Dec 12, 2006, at 4:11 AM, McBride, Brian wrote:
> I note the result [3] contains
>
> [[
>   <rdf:Description rdf:about="">
>     <transformation xmlns="http://www.w3.org/1999/xhtml"
> rdf:resource="RDFa2RDFXML.xsl"/>
>   </rdf:Description>
> ]]
>
> The rdf:about refers to the RDF document, which from my reading of the
> latest (editors draft) RDF/A spec [4] is incorrect - see 4.3.3.1.  It
> should refer to the source document.

Odd; I would have thought that sort of thing would show up when I ran 
the
tests. I wonder if the test harness does something there... i.e. treats
the output document as a representation (a la a cached copy) of the
source URI.

> I note also that the property in this statement has a URI of
> http://www.w3.org/1999/xhtmltransformation and is affected by the
> default namespace in effect in the head of the source document.  I 
> think
> the test case accurately reflects the spec, but that still looks a bit
> broken to me.

Indeed, that is odd. Fabien?

> The statement
>
> [[
>   <rdf:Description
> rdf:about="http://www.w3.org/TR/2004/REC-rdf-mt-20040210/">
>     <dc:title xmlns:dc="http://purl.org/dc/elements/1.1/">RDF Semantics
> - W3C Recommendation 10 February 2004</dc:title>
>   </rdf:Description>
> ]]
>
> The RDF/A spec says the object of this property should be an XML 
> literal
> (see 4.4.1) and the RDF spec [5] says that XML literals and plain
> literals are disjoint.  I think this may be a problem with the RDF/A
> spec.
>
> I've run out of time - too many specs to read, but there is another
> issue that I think I ought to raise.  (Must you - do I hear?)
>
> I'd like to see a separation between test cases and examples.  I'd
> prefer test cases that were small, simple, verifiable by inspection and
> didn't carry any baggage such as "is this a correct interpretation of
> the RDF/A spec".  Our test cases are going beyond what the spec 
> defines,
> e.g. in illustrating practice for XSLT transforms - and Dan asked
> recently whether they should be normative.  Is it then a good idea to
> include an RDF/A transform as a test case?

Small, simple test cases are more than welcome. The best way
to show what you'd like is to contribute it. (Thanks for the base URI
tests; I intend to pore over them in detail shortly).

I see a lot of value in test cases that integrate real-world examples,
so that's what I intend to contribute.
The fact that you found a quirk around 
http://www.w3.org/1999/xhtmltransformation
is an important part of the process. I think it was a bit above and
beyond the call of duty to read the RDFa spec, but I'm not sorry
we got you to take a look.

> Its relatively
> straighforward to have a different class called examples - though some
> work for someone to set it up.

I have done some work in that direction; note the 
"Example/advocacy/application test ideas"
section heading in the test overview.
   http://www.w3.org/2001/sw/grddl-wg/td/testlist1

I was thinking that tests would migrate out of there when they 
transition from ideas
to actual tests, but I'm OK with maintaining a separation.


> Brian
>
>
> [1] http://www.w3.org/2006/12/06-grddl-wg-irc
> [2] http://www.w3.org/2001/sw/grddl-wg/td/testlist1#rdfa1
> [3] http://www.w3.org/2001/sw/grddl-wg/td/rdf_sem-output.rdf
> [4] http://www.w3.org/2006/07/SWD/RDFa/syntax/
> [5] http://www.w3.org/TR/rdf-concepts/#section-XMLLiteral
>


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 12 December 2006 16:59:17 UTC