RE: Interpreting RDF as representation of a resource?

Hi Jan,

This is an issue that we have bumped into as well.

However - I would like to characterise the problem slightly differently. For me it is not so much that you "lose the typing information" - I think the types are still there (either explicitly or implicitly) - the problem seems to be how do we characterise what is the intended "major" node within the graph.

As you say with XML it's easy because it is ordered - the root node is presumed to represent the major node. But with RDF it's tricky.

The best we can come up with at the moment is to attach a property that is unique within the graph being posted that allows us to say "this is the major node". It's not a very generalised solution though.

Cheers,

Murray

>-----Original Message-----
>From: www-rdf-interest-request@w3.org 
>[mailto:www-rdf-interest-request@w3.org] On Behalf Of Jan Algermissen
>Sent: Wednesday,24 November 2004 5:59 AM
>To: Mark Baker
>Cc: Patrick.Stickler@nokia.com; www-rdf-interest@w3.org
>Subject: Re: Interpreting RDF as representation of a resource?
>
>
>Mark Baker wrote:
>
>> One of us misunderstood Jan, it seems. 8-/
>
>my question was actually more targetted at RDF as opposed to the 
>HTTP request issues. 
>
>Given an XML document, I know (assuming it references some 
>kind of schema)
>what the type of the entity is that it represents (e.g. 
>purchase order)[1]
>
>If designed well, I can interprete the same document as RDF 
>(gaining all
>the wonderfull merging and partial understanding  possibilities). OTH,
>given the RDF I completely loose the typing information which I'd like
>to use for further dispatching (e.g. selecting the appropriate 
>stylesheet
>for rendition).
>
>Any thoughts on how to obtain the typing information from the RDF?
>
>Jan
>
>[1] I'd really like to expand this whole issue towards Duck Typing
>( http://www.propylon.com/news/ctoarticles/040224_duckmodeling.html ),
>but simple things first :o)
>
>

Received on Tuesday, 23 November 2004 23:39:00 UTC