- From: Pierre-Antoine CHAMPIN <champin@cpe.fr>
- Date: Tue, 23 Nov 1999 09:27:47 +0000
- To: seth@halcyon.com
- CC: RDF Intrest Group <www-rdf-interest@w3.org>
> Pierre-Antoine CHAMPIN wrote:
> > where "thisfile.rdf#genid1" whould have "This is a litteral" as CONTENT...
> > which surely would raise some implementation issues left to discuss!
I knew it would :)
Seth Russell wrote:
> But suppose then we want to say something about "this literal" ... would we
> then write:
>
> <rdf:Description about="This is a litteral">
> <s:propB> Wanna take a bath</s:propB>
> </rdf:Description>
>
> ___OR___
>
> <rdf:Description about="thisfile.rdf#genid1">
> <s:propB>Wanna take a bath</s:propB>
> </rdf:Description>
The rdf:about attribute has to receive an URI,
so only the second proposition looks correct to me.
A problem is :
it makes the assumption that any RDF parser will generate
IDs for anonymous resources in the same way,
which is a strong hypothesis.
But I think Sergey Melnik initiated a thread about ID uniqueness on this list.
Actually, directly expressing statements about anonymous resources
is never easy in RDF. The better way would be to put them in a bag,
and then assert an aboutEach description.
<rdf:Bag ID="anonymousThings">
<rdf:li> This is a literal </rdf:li>
<rdf:li parseType="Resource">
<s:p1> This is an </s:p1>
<s:p2> anonymous Resource </s:p2>
</rdf:li>
</rdf:Bag>
<rdf:Description aboutEach="#anonymousThings">
<s:p3> A prop for anonymous resources </s:p3>
</rdf:Description>
Pierre-Antoine
Received on Tuesday, 23 November 1999 03:32:30 UTC