[topicmapmail] multiple Source Locators

[Thomas B. Passin]

Lars Marius Garshol wrote:

> | [how to define semantics of URI references in RDF]
> | 
> | Right, so you do it by using them in a context defined by a
> | predicate.
> 
> Actually, I'm not convinced that is true. It would be really nice if
> it were, but my impression is that seeing
> 
>   (uri1, uri2, <whatever>)
> 
> doesn't necessarily tell you how to interpret uri1. Sometimes the
> rdfs:domain of the property would tell you, but sometimes the
> rdfs:domain is rdfs:Resource and then you have no idea.
> 

uri1 is an abstract identifier.  You can learn nothing from it by 
itself.  It is true that the rdfs:domain of a property would tell you 
the class of the resource *if*

1) the domain were defined to include just a single class, and
2) Someone did not use the property in violation of the triple stating 
its domain.

In any case, it took at least one other triple to convey the 
information, so once again "uri1" by itself conveys nothing except the 
identity of a specific node in the rdf graph.  Well, yes, its very 
presence in a statement implies that there is a resource having that 
identifier.  Thus, an rdf statement acts like the infamous "topic demander".

Cheers,

Tom P

-- 
Thomas B. Passin
Explorer's Guide to the Semantic Web (Manning Books)
http://www.manning.com/catalog/view.php?book=passin