Re: Integrity constraints in DAML-O (was: Chainsaw?)

Seth,

At 12:02 PM 11/3/00 -0800, Seth Russell wrote:
>Graham Klyne wrote:
>
> > I think a "proper name" construct can be introduced without extending or
> > bending RDF as something like this:
> >
> >     [AnyResource]--properName-->[(anon?)] --rdf:Type--> [ProperName]
> >                                 [       ] --value-----> "(name string)"
>
>Can I ask a really stupid question:  Why can't this just be:
>
>[AnyResource]--properName-->"(name string)"
>[AnyResource]--rdf:Type-->[ProperlyNamed]

It's obviously not a stupid question because I can't remember the answer!

What you propose was my first idea, but then I had a problem with it.

Vaguely, I think that I wanted the name itself to be a resource so I could 
use it as a subject and/or object of other RDF statements.  I'll try and be 
more specific when I recall the exact reason.

>What is the utility of the extra node?  Can't the schema for "properName" 
>indicate that
>its object is expected to be a public string just like "rdf:value"?  Also 
>can't
>{[AnyResource]--rdf:Type-->[ProperlyNamed]} be inferred by relationships 
>hung off of
>properName ?

Now there's an interesting thought I hadn't considered.  There may be other 
hidden possibilities here.

#g
--

[later]

Aha!  I've just remembered why I put in the extra node.

I felt that it was important that the proper name Boston referring to the 
east coast US city should be distinguishable from the same proper name 
referring to a small English town, etc.  I couldn't see how that 
distinction could be drawn using just literals.

However, your idea of hanging additional info off the properName property 
arc (or, strictly, off the reification of the properName statement) is, I 
think, truer to the concept here.  The proper name string "Boston" is after 
all just a string.  It's the way it's used (the entities to which it is 
applied) that distinguish its different uses.

BTW, I don't think the ProperlyNamed class is really needed, since the same 
information can be inferred by the presence of the properName property 
(...maybe).

Thank you for your intervention ... that was helpful.

#g

------------
Graham Klyne
(GK@ACM.ORG)

Received on Saturday, 4 November 2000 07:26:26 UTC