Re: two comments on RDF Schema spec

Dan Brickley wrote:
> Guus,
> 
> I've switched this from www-archive to www-rdf-comments, to register it 
> as a Last Call comment. For detailed discussion lets take it back to 
> www-archive.

Sorry about this. I didn't actually want to make a formal LC comment; it 
were more like two points where I wanted to check my understanding. 
These came up when we upgraded our tools to the new RDF spec (so really 
outisde the scope of my Webont co-chair role).
I only use www-archive to have a reference, if needed in the future.

See below for specific points.

> 
> * Guus Schreiber <schreiber@swi.psy.uva.nl> [2003-01-27 15:33+0100]
> 
>>Dan,
>>
>>Two small commewnts on the RDF Schema spec:
>>
>>
>>>2.3 rdfs:Literal
>>>
>>>..
>>>
>>>rdfs:Literal is a subclass of rdfs:Resource.
>>
>>Is this correct? I thought rdfs:Literal was not a resource.
>>This is also what rdfs.rdfs states.
> 
> 
> I believe "rdfs:Literal is a subclass of rdfs:Resource" to be 
> the current wisdom of the Semantics spec. 
> 
> I'm not sure what you mean by rdfs.rdfs. If you mean the 
> RDF schema at the RDFS namespace URI, that document is in need of 
> updating. If you mean the RDF schema for RDF schema that is 
> part of the LC working draft, we may have an issue.

With rdfs.rdfs I meant the "RDF Schema in RDF/XML" in Appendix A:

<rdfs:Class rdf:about="http://www.w3.org/2000/01/rdf-schema#Literal">
   <rdfs:isDefinedBy
rdf:resource="http://www.w3.org/2000/01/rdf-schema#"/>
   <rdfs:label>Literal</rdfs:label>
   <rdfs:comment>The class of literal values, eg. textual strings and 
integers.</rdfs:comment>
</rdfs:Class>

If I understand you correctly, a "subClassOf Resource" statement should 
be added to this definition of Literal. We noticed it because it now 
turns up as an "orpajn" node in our RDFS subclass tree.

My confusion came from my intuition of resources and literals being 
disjoint. Having reread section 3.3.1 in the Semantics doc ("A note 
anout rdfs:Literal"), I now understand the spec. It could be worthwhile 
to add a explanatory note in the Schema document, because I guess other 
people might get confused in the same way.

> 
>>>3.2 rdfs:domain
>>>
>>>[..] >
>>
>>>The rdfs:range of rdfs:domain is the class rdfs:Class. 
>>>This states that any resource that is the value of an
>>>rdfs:domain property is an instance of rdfs:Class.
>>
>>This leaves open the possibility that rdfs;Literal
>>(an instance of rdfs:Class)  can be defined as a domain.
>>Is this intended?
> 
> 
> If there were a true statement that 
> some property p1 had an rdfs:domain of rdfs:Literal, we would 
> have a problem. I believe the view is that there are no such true 
> statements. 

My confusion stemmed from the same source, namely that literals cannot 
be objects in RDF statements, thus allowing them as a rdfs;domain  is 
counterintuitive. Again, the note in Semantics explains this.

> If there were a class defined for resources that aren't literals, 
> perhaps we could express this more clearly in machine-friendly terms. Does 
> OWL define such a class? 

Not really. The notion of Individual in OWL is much restrictve. You 
could of course define a subclass of Resource such as 
NonLiteralResource, but this would involve careful checking of the 
metamodel.

Guus

> Dan
> 
> 

-- 
A. Th. Schreiber, SWI, University of Amsterdam,
http://www.swi.psy.uva.nl/usr/Schreiber/home.html

Received on Tuesday, 28 January 2003 09:01:20 UTC