Re: bNodes wanted

> Two cases where I use bNodes (which may be different to what they are meant
> for):
> 
> 1/ the thing is not web resource (people, organisations, email messages and
> that Porsche I don't have)
> 2/ grouping to build information about a composite concept
> 
> Abusing the ontology for ISWC:
> 
>   <Researcher>
>      <homepage rdf:resource='http://www-uk.hpl.hp.com/people/afs'/>
>      <name rdf:parseType="Resource">
>         <first>Andy</first>
>         <last>Seaborne</last>
>      </name>
>   </Researcher>
> 
> I am not a web resource but can be found by my homepage.  My name has
> structure and this could be useful to retain.
> 
> You can refer to a bNode - you find it by query.  It is especially
> interesting if you find more than one.  URIs don't really have such a
> priviledged place - we could have all bNodes and a property "hasURI" and
> then everything is found by query.

These are nice places to use bNodes, but you could use a URI just fine
here.  You just make up some otherwise-unused URI (Skolemize the
bNode) and things are fine.   The only time you change the meaning of
RDF via Skolemization is when the RDF is not asserted.   As I think
you pointed out it is nice to know no-one else has the node, but
that's just an engineering convenince.

I pointed out a few weeks ago that the hasURI property (which I called
uname, since its important feature is that it's an unambiguous
property, and TimBL calls log:uri) might reasonably be equated to
rdf:about; doing so is a pure extension of M&S 1.0 and simplifies the
whole RDF model *and* syntax.

    -- sandro

Received on Friday, 24 May 2002 11:49:15 UTC