W3C home > Mailing lists > Public > public-grddl-wg@w3.org > April 2007

Re: hotel stuff from primer as test? no

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Fri, 20 Apr 2007 21:36:09 +0100
Message-ID: <46292439.8020602@hpl.hp.com>
To: Harry Halpin <hhalpin@ibiblio.org>
CC: GRDDL Working Group <public-grddl-wg@w3.org>




A little of the output is:

    <Description xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#" 
rdf:about="_593">
       <vcard:FN>Elena Plaza Hotel</vcard:FN>
       <type xmlns=""/>
       <region xmlns=""/>
       <locality xmlns="">Helsinki</locality>
       <country-name xmlns="">Finland</country-name>
       <postal-code xmlns="">FIN-00100</postal-code>
       <street-address xmlns="">Kikonkatu 23</street-address>
       <extended-address xmlns=""/>
       <post-office-box xmlns=""/>
       <review:hasReview>
          <review:Review>
             <dc:title>Elena Plaza Hotel</dc:title>
             <dc:type>business</dc:type>
             <review:createdOn>20050607T211457Z</review:createdOn>
             <review:rating>5</review:rating>
             <review:reviewer>
                <foaf:Person>
                   <foaf:name>JohnD</foaf:name>
                   <foaf:homepage rdf:resource="http://john.example.org/"/>
                </foaf:Person>
             </review:reviewer>
          </review:Review>
       </review:hasReview>
    </Description>

Many of the properties have xmlns="" on them.
A buggy xslt implementation might omit these in error.
With them, which is the correct result from the transform, this is not 
valid RDF/XML.

This appear to come from
http://suda.co.uk/projects/microformats/mf-templates.xsl
which is in imported from the main transform. This in turns imports 
other stylesheets. It's a lot of code. Too much in my view for a test 
case - particularly too much code for a broken test case when we have 
only a few days.

Jeremy



Harry Halpin wrote:
> 
> Could  you tell me exactly what of the transforms doesn't work and 
> produces invalid RDF/XML (I tend to use xalan)? It's very  likely we 
> checked in (for example, I know this is the case with RDFa) and linked 
> to out-of-date XSLTs?
> 
> I'll work on fixing this up over the weekend...
> 
> 
>  On Fri, 20 Apr 2007, Jeremy Carroll wrote:
> 
>>
>>
>> I don't think the hotel data stuff is suitable as a test:
>>
>> - it's too complicated
>> - the imports closure is quite large, four or five files maybe - I 
>> haven't found all of them yet; for a test we would need to copy all of 
>> them
>> - the current transforms are broken and too big to fix easily
>>  [problem is that the RDF/XML output does not validate, because some 
>> elements are unqualified - a buggy XSLT implementation may qualify 
>> them by mistake]
>>
>> It's also highly desirable to address some of these issues before next 
>> publication of the primer.
>>
>> Jeremy
>>
>>
>>
> 

-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England
Received on Friday, 20 April 2007 20:38:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:48 GMT