W3C home > Mailing lists > Public > www-rdf-comments@w3.org > July to September 2002

Re: RDF Issue rdfs-clarify-subClass-and-instance

From: graham wideman <graham@wideman-one.com>
Date: Wed, 28 Aug 2002 12:49:10 -0700
Message-Id: <4.1.20020828121331.017489b8@sunstroke.sdsu.edu>
To: Graham Klyne <GK@ninebynine.org>
Cc: www-rdf-comments@w3.org

Graham (nice name, by the way :-)

At 8/28/2002 04:30 PM, Graham Klyne wrote:
>At 07:13 AM 8/28/02 -0700, graham wideman wrote:
>>Summary:
>>Old: Class membership implies something about properties of instances
>>New: Properties of instance imply class membership
>
>Roughly, yes (to the 'New:').
>
>I don't think there's been a recent change of understanding, even if the 
>words themselves have changed recently.

I contrasted Primer 2002-04-26 section 4.2 versus Primer 2002-08-23 section 5.2, and to me they looked pretty much incompatible with one another, which I infer means that someone's understanding changed quite a bit:

2002-04-26 S4.2: "A domain constraint specifies that a property may be used on resources of a certain class".  

2002-08-23 S5.2: "If ex:weight has [a] domain property specifying ex:Book as the domain,  this is a statement that any resource that has a ex:weight property is an instance of class ex:Book.

>  RDF has always espoused the idea 
>that "anyone can say anything about anything", against which background 
>it's difficult to make sense of RDFS as constraints without nobbling its 
>descriptive role.
>
>So RDF Schema does not, of itself, limit what one can say about instances 
>of classes mentioned in the schema

But my point was, that according to the Primer (which may of course be wrong), RDFS *does* limit what you can say about instances of classes mentioned in the schema, and severely so. It prescribes the meaning of the rdfs:domain is such a way that (so far as I can see):

a) It *prevents* employing rdfs:domain to specify the mundane kind of constraint that one frequently needs (requiring particular properties on instances of particular classes), and 

b) It *prescribes* a requirement that RDFS users treat membership in particular classes as a side-effect of each property it possesses, considered individually!  I might even buy a scheme in which class membership resulted from considering the *total* of all properties (4 legs AND furry AND barks... ah-ha, it's a dog!), but the RDF scheme described in the Primer is arbitrarily at odds with the world that RDF is supposed to describe.

If that's not "limiting what one can say" relative to applying rdfs:domain, I don't know what is!

Regards,

Graham

---------------------------------------------------
Graham Wideman
Resources for programmable diagramming at:
http://www.diagramantics.com
graham@wideman-one.com
http://www.wideman-one.com
Received on Wednesday, 28 August 2002 15:51:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:30 GMT