Re: The URI of a RDDL "nature"

On Jan 18, 2006, at 4:48 AM, Williams, Stuart ((HP Labs, Bristol))  

> Hello Norm,
> I'm just catching up on this thread...
>> It is with this in mind that the TAG wonders if you'd be
>> willing to establish new URIs with the pattern
>> for the natures. I would suggest preserving, but deprecating,
>> the natures listed above (so that there would be two natures for  
>> those
>> resources) and simply dropping the rest.
> Under this proposal, RDDL natures become a closed space under the
> control of the maintainer/owner of rather than an openly
> extensible space where anyone could contribute a new nature. Is that
> really what the TAG wants?

I hope not. I don't want to speak for the TAG, but the intention of  
both the RDDL properties and RDDL natures documents was to give  
examples and jumpstart the process so that other people could define  
their own natures and properties. What has happened is that the  
documents have become popular and I get requests from time to time to  
add natures and properties which have general utility. The documents  
hadn't been edited  in a few years actually, until the TAG has  
recently reviewed them. Upon reflecting, these suggestions are really  
good and particularly the natures document needs some fairly  
significant revision. I am not volunteering to maintain the RDDL  
equivalent of CYC :-))
> It strikes me that one would want the space of rddl natures to be  
> openly
> extensible so that one could say something like:
> 	:myNatureTerm a rddl:nature .
> in order to declare term as a rddl nature. Once could of course the
> dereference the nature term to find more (machine/human) readable
> documentation about that nature (maybe in another namespace document
> :-).

Yes. The latest provides: defines:

=> owl:subClassOf owl:Class

as well as

 > owl:subClassOf owl:Class

and that we *could* go so far as to say owl:disjointWith http://

It should also be noted that one can define a class of arbitrary  
complexity "in" RDDL simply by creating a link and then defining an  
owl:equivalentClass which points toward an actual OWL/RDFS definition  

<a id="FooClass" rddl:purpose="owl:equivalentClass" href="http://"> ...



Received on Wednesday, 18 January 2006 16:20:04 UTC