Re: RDF Structure (was Extension property: "using")

Hi Charles and Chris,

This is our conclusion as well, after having played a bit with EARL and
RDF, that EARL may be created by simpler XML tools that are not RDF
aware, as long as the RDF graph is correct. 

However when operating on all the different XML variants of EARL reports
we may get in, then we have to use an RDF library to focus on the graph
structure instead of XML syntax, and also maybe some kind of RDF query
languages like e.g. SPARQL.


tor, 14,.04.2005 kl. 00.34 +1000, skrev Charles McCathieNevile:
> On Wed, 13 Apr 2005 18:16:33 +1000, Chris Ridpath  
> <> wrote:
> > This is one topic that I'm still not clear on so please forgive my  
> > ignorance. I assume that our EARL, which is RDF and XML, will have a  
> > schema which states which elements and attributes can be present. The  
> > schema may also state the order in which some elements must occur. It  
> > may also state which elements may contain other elements. Are my  
> > assumptions correct?
> Nope.
> > I know that you can mix RDF documents but my concern is that it will be  
> > difficult to process the EARL as XML because I won't know where things  
> > are.
> Right. Unless you produced it yourself, you cannot expect it to always  
> look the same.
> > For example, the file that passed or failed a guideline will always be  
> > marked like:
> > <earl:WebContent rdf:about="#subject">
> >  <earl:reprOf rdf:resource="URI to the file"/>
> > </earl:WebContent>
> >
> > Is that right?
> No. It could equally be
> <r:Description r:ID="subject">
>    <r:type r:resource="&earl;WebContent" />
> </r:Description>
> <r:Description r:about="#subject">
>    <e:reprOf r:resource="URI to the file" />
> </r:Description>
> which, in RDF, means *exactly* the same thing.
> If we have a suggested XML schema - i.e. one that lines up elements and  
> attributes in a particular way, for some subset of EARL information, then  
> we will need to write a SPARQL query that constructs things in this way. I  
> am not convinced that this is a worthwhile exercise, since processing as  
> XML places serious limits on what is actually interoperable (these are  
> being reduced slowly, but will continue to exist for the future) and  
> encourages people to regard the strictly limited XML version as equal.
> It also means we need to do a lot more work to define the things we want,  
> since the XML version will not have the flexibility and mergeability of  
> RDF (unless people dumb down to only using well-formed XML, in which case  
> semantic interoperability just about vanishes).
> cheers
> Chaals
Nils Ulltveit-Moe <>

Received on Wednesday, 13 April 2005 15:15:48 UTC