Re: Dave Beckett's comment concerning base-uri

Unless I made a mistake, if no explicit base and no parameters have been 
specified, relative URIs are used throughout the produced RDF/XML.

Fabien


Chimezie Ogbuji a écrit :
> On Tue, 2007-05-08 at 14:43 +0100, Jeremy Carroll wrote:
>   
>> I note
>>
>> http://lists.w3.org/Archives/Public/public-grddl-comments/2007AprJun/0051.html
>>
>> with some concern.
>>
>> The GRDDL spec does not specify any parameters being passed to the XSLT. 
>> My implementation does not do so, and I don't see any easy way of 
>> deciding which parameters to pass, in the absence of a normative list.
>>     
>
> Yes, this is the biggest concern I have with specifying a parameter:
> what do you name it? If you decide on a name doesn't that essentially
> become a 'reserved' parameter for all GRDDL transforms - seems a bit
> heavy-handed to me.
>
>   
>> As far as I can tell, Dave's comment concerns particular examples; in as 
>> much as these examples are WG output we should fix them.
>>     
>
> I agree, we should be sure that any test case / example / transform we
> produce should deal with this without the need for an explicit
> parameter.
>
>   
>> My understanding is that our resolution of the base URI issue allows 
>> XSLT 1.0 transforms not to worry excessively about base URIs since 
>> relative URIs from the content can be copied over into the RDF/XML, 
>> where they will be resolved against the document base URI. 
>>     
>
> In addition XSLT 1.0 transforms also have the option to 'reap' for
> xml:base attribute (as well as XHTML Base elements) explicitly.  This is
> what RDFa2RDFXML.xslt does.
>
>   
>> Our examples 
>> should demonstrate this.
>>     
>
> The one example he mentions explicitly seems to already demonstrate some of this:
>
> <!-- uri of the current XHTML page -->
> <xsl:param name="this" select="//*/@xml:base[position()=1]"/>
>
> Ofcourse, it could also use relative URIs throughout instead of attempting to resolve them explicitly against the base it reaped from content.
>
> I.e., instead of:
>
> <when test="parent::*/attribute::id">
>   <value-of select="concat($this,'#',parent::*/attribute::id)"/>
> </when>
>
> Do:
>
> <when test="parent::*/attribute::id">
>   <value-of select="parent::*/attribute::id"/>
> </when>
>
> And allow the RDF processor which parses the RDF/XML to handle the URI resolution against a given base - as you suggest.
>
>   

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

Received on Tuesday, 8 May 2007 17:15:37 UTC