Re: ANNOUNCEMENT: RDFStyles: alternative to XSLT for RDF

Libby Miller wrote:
> Indeed, but RDFPath-type languages model the 'path' in terms of the RDF
> model, not the XML model. They steal the syntax to mean something
> different. Such a system could be used with XSLT to query the RDF model.

Agreed. That's why I was making the difference between XPath and XSLT. 
XPath-based RDFPath is fine with me. What I am questioning is the direct 
processing of RDF through RDF/XML + XPath/XSLT as they stand today.

> If RDF had a cannonical syntax in XML then you could use XMLPath with
> RDF and although you would not be querying the RDF model, it would
> nevertheless work ok.
> 
> Libby

I also fully agree with that. But we don't (yet?) have a cannonical 
RDF/XML syntax. So it is not ok right now. And although it is ok from a 
pragmatic point of view, it is not from a more conceptual one (in my 
opinion). We can probably live with that, but we should be careful as it 
might not be in our best interest on the long run.

Emmanuel

> 
> On Thu, 23 Oct 2003, Emmanuel Pietriga wrote:
> 
> 
>>>>XSLT is a pretty good hammer. But why does anyone think that XSLT is the
>>>>right general-purpose tool for manipulating RDF information?  Beats me.
>>
>>I don't think it is. More precisely, I don't think XPath (and thus XSLT
>>as it exists now) should be used to process RDF, because this means
>>addressing the problem at the wrong level:
>>
>>When we use XPath to query an XML document, XPath is concerned with the
>>  XML structure itself, i.e. the XML tree, not its serialization or what
>>the tree represents at a higher level of abstraction (in the case of
>>RDF, a directed graph).
>>
>>RDF/XML is just a serialization of RDF. So when we make XPath queries on
>>an RDF/XML document, we do not query the RDF graph (which is what we
>>want to do), but a projection of this graph onto an XML tree (which is
>>the cause of the problem of the multiple representations of the same
>>graph in RDF/XML).
>>
>>As many other people, I believe it would be nice to be able to use
>>existing processing tools like XSLT with RDF, as it would make things
>>easier and reduce the cost of learning and manipulating RDF. But some
>>technologies are just not appropriate, like XPath. So we need some kind
>>of RDFPath language that addresses the problem of selecting parts of an
>>RDF model at the appropriate level (which is not an XML tree
>>representation of the RDF graph built from its RDF/XML representation)



-- 
Emmanuel Pietriga (epietriga@nuxeo.com)
tel (mobile): +33 6 88 51 94 98
http://claribole.net

Received on Thursday, 23 October 2003 09:47:45 UTC