GRDDL RDFa transformation tests results

Hello,

I had a look at the test cases and here is a summary of what I've done 
this week-end:
- I released v0.7 of the transformation stylesheet:
http://www-sop.inria.fr/acacia/soft/sweetwiki.html

- I built a JUnit test suite to automate the test of the output of the 
transformation using the SPARQL engine CORESE
http://www-sop.inria.fr/acacia/soft/corese/

- I added 3 tests to the current collection (multiple rel / property 
predicates and rel with no href):
http://www-sop.inria.fr/acacia/personnel/Fabien.Gandon/tmp/grddl/rdfatests/

- Among the 32 tests I pass all of them but 5 : I fail 1 of them, I 
think 4 of them have mistakes and 1 succeeds but I am not sure I like 
what it does. Results of GRDDL transformations and SPARQL queries are 
available here:
http://www-sop.inria.fr/acacia/personnel/Fabien.Gandon/tmp/grddl/rdfatests/results_2007_04_15/

Details below,

Cheers,



Details:

Question #1: Where do I get the URL of the source document from?
The current stylesheet tests if there is a xml:base. If there is no 
xml:base it tests if the parameter URL was past to it. If not the 
default value is "".


Question #2: What happened to the use of "class" attribute?
Since I found no trace of it in the tests and in the syntax, I 
deactivated its use in the transformation.


Question #3: What should I do with non qualified predicates?
Ex: <link rel="transformation" 
href="http://www.w3.org/2002/12/cal/glean-hcal"/>


Test case#11: I think the SPARQL query is not in the right direction.
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000011.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000011.sparql

The RDFa says (note that it uses "rev" and not "rel")
<link about="http://example.org/Person1"
          rev="foaf:knows" href="http://example.org/Person2" />

While the query says:
<http://example.org/Person1> <http://xmlns.com/foaf/0.1/knows> 
<http://example.org/Person2>

Even if foaf:knows was transitive, I think it wouldn't be a good idea to 
rely on this in an RDFa test.


Test case#13: I fail this test but I need explanations.
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000013.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000013.sparql
Here I don't really know what to produce: the reason I fail is because 
when I copy the source nodes:
E = mc<sup>2</sup>: The Most Urgent Problem of Our Time
The transformation adds the namespaces declaration to <sup>.
I am not clear exactly what should be copied: (1) the exact string or 
(2) the XML nodes and their context?
I looked at the following pointer but couldn't tell after either:
http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#dfn-rdf-XMLLiteral


Test case#102: the SPARQL query contains "malto:libby" instead of 
"mailto:libby".
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000102.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000102.sparql


Test case#105: the SPARQL query contains "000105.ttl#photo1" instead of 
"000105.html#photo1".
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000105.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000105.sparql


Test case#108: the SPARQL query contains several "000108.ttl#me" instead 
of "000108.html#me".
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000108.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000108.sparql


Test case#112: Although I pass this one I am not sure it is right to use 
datatype like this.
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000112.html
http://www.w3.org/2006/07/SWD/RDFa/testsuite/testcases/000112.sparql
This one uses a fake datatype which implies hacking in the code and I am 
not sure it is a good idea. Why not use rdf:Literal for the same result?
<span about="http://example.org/node" property="ex:property" 
xml:lang="fr" datatype="plainliteral">chat</span>




Hausenblas, Michael:
> Fabien,
>
> Ok, great. Let us go through it step by step.
>
> First, my action [1] (in the RDFa TF) was about
> updating your XSLT to reflect the latest updates,
> for example [2].
>
> Regarding the URL for the GRDDL profile, I'd like propose
> to take Dan's comment [3] into account.
>
> I'm in doubt if it makes sense that I actually start making
> the updates, as you already have planed to do it ...
>
> Regarding the test cases: The RDFa Test Suite is available
> at [4] and we would be very happy if people start using it and
> give feedback ;)
>
> IMHO it would be a good idea, if you could join us at next 
> Monday's RDFa telecon. We could sort out these issues -
> also w.r.t. overlaps in the GRDDL/RDFa - and agree on a 
> further road map.
>
> Now, what do you think?
>
> Cheers,
> 	Michael
>
> [1] http://www.w3.org/2007/04/11-rdfa-minutes.html#action12 
> [2] http://www.w3.org/2006/07/SWD/wiki/RDFa/LiteralObject
> [3]
> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Apr/0027.
> html
> [4] http://www.w3.org/2006/07/SWD/RDFa/testsuite/
>
> ----------------------------------------------------------
>  Michael Hausenblas, MSc.
>  Institute of Information Systems & Information Management
>  JOANNEUM RESEARCH Forschungsgesellschaft mbH
>  
>  http://www.joanneum.at/iis/
> ----------------------------------------------------------
>
>
>   
>> -----Original Message-----
>> From: Fabien Gandon [mailto:Fabien.Gandon@sophia.inria.fr]
>> Sent: Friday, April 13, 2007 10:29 AM
>> To: Hausenblas, Michael
>> Cc: public-rdf-in-xhtml-tf@w3.org; public-grddl-wg; Dan
>> Connolly; Harry Halpin
>> Subject: Re: Comments on RDFa Syntax version 1.8 2007/04/06 16:36:34
>>
>> Hausenblas, Michael:
>>
>>
>>       Fabien,
>>        
>>
>>               I hope to get time to update the RDFa2RDFXML
>> transformation [2] to
>>               reflect latest changes.
>>                  
>>
>>       What a coincidence! I got an action on exactly this one -
>>       we may want to jointly do that?
>>       This also goes along with setting up an official GRDDL
>> profile for RDFa.
>>        
>>
>>
>> Hi Michael,
>>
>> I didn't know there was an action on this and now that I
>> finished the GRDDL use case note, I intend to update the RDFa
>> transformation again. In fact this is the 6th version of the
>> RDFa2RDFXML stylesheet and here is what I had on my "to do list":
>> 1 - INRIA accepted to set up a permanent domain name and
>> server to publish resources with stable durable URL. As soon
>> as it is running I intend to use it to publish a profile and a
>> transform with stable URL.
>> 2 - The last version of the transform addresses a number of
>> features (e.g. CURIEs) but a number of cases are still not
>> addressed yet (e.g. multiple values in "rel" attribute).
>> 3 - There is also a need to produce a test suite to validate
>> the transformation. For the time being I use a file compiling
>> some of the examples of the primer but the tests need to be
>> complete w.r.t. last specs.
>> 4 - There is a need to update the GRDDL+RDFa primer I started
>> a while ago and which is out of date now ; see 
>> http://www-sop.inria.fr/acacia/personnel/Fabien.Gandon/tmp/grdd
>>     
> l/rdfaprimer/PrimerRDFaSection.html
>   
>> Cheers,
>>
>> --
>> Fabien - http://www.inria.fr/acacia/fabien/
>>
>>     
>
>   

-- 
Fabien - http://www.inria.fr/acacia/fabien/