W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2001

RE: RDFCore Update

From: <Patrick.Stickler@nokia.com>
Date: Fri, 19 Oct 2001 20:08:20 +0300
Message-ID: <2BF0AD29BC31FE46B78877321144043114C038@trebe003.NOE.Nokia.com>
To: geoff@sover.net, www-rdf-interest@w3.org


> -----Original Message-----
> From: ext Geoff Chappell [mailto:geoff@sover.net]
> Sent: 19 October, 2001 18:39
> To: Stickler Patrick (NRC/Tampere); www-rdf-interest@w3.org
> Subject: Re: RDFCore Update
> 
> 
> 
> ----- Original Message -----
> From: <Patrick.Stickler@nokia.com>
> To: <geoff@sover.net>; <bwm@hplb.hpl.hp.com>; 
> <www-rdf-interest@w3.org>
> Sent: 10/19/2001 11:06 AM
> Subject: RE: RDFCore Update
> 
> 
> >
> >
> > Consider a piece of fruit. One may define for a property 
> something like
> > 'rdfs:range="#fruit"' and then one could assign various 
> values having
> > types such as "#orange", "#banana", "#kiwi", etc., all of which are
> > sub-types of "#fruit" and an RDFS schema validator is happy with any
> > of them, because they are all valid "#fruit" values -- but 
> nevertheless
> > you *still* have to inform the RDFS validator the exact 
> type for each
> > value. You can't derive it from the rdfs:range, as that is a general
> > type, not the *actual* type -- and you need to know the 
> actual type, as
> > it can affect system behavior (one might do things with 
> bananas that one
> > would never do with pineapples ;-) and you wouldn't want to have to
> > enumerate all the descendant types of "#fruit" as ranges, as that
> > precludes flexible extensibility (the ability to define new 
> types of fruit
> > which are valid values for that property).
> >
> 
> I guess I don't see this as an argument for embedding type in 
> a uri (unless
> I've misunderstood you). What if you wanted to further 
> classify and object
> into an even more specific class? you'd have to give it a new uri?

You have misunderstood the nature of a URV. Please see my
overview of this concept posted earlier to the list, in case
you haven't read it yet. It sets the context for most of my
discussion:

http://lists.w3.org/Archives/Public/www-rdf-comments/2001OctDec/0088.html

A URV is comprised of two components, the data type and the
literal value. It is not meant to denote any other resource. If you
change the data type of a literal value, you change the way
it is represented as a URV, hence, yes you change the URI, but
that's not like changing the URI of e.g. a property name or
a web page.

Thus, one may in one context define 'xyz:integer:5' and in
another context 'xyz:count:5'. Are those the same literal values?
Maybe. Maybe not. They certainly are not of the same data type.

So if you change the data type of a literal, yes the URV (URI)
will change. But that is no different really than having
two qualified anonymous nodes with the same literal and different
rdf:type's specified. No?

> One solution is to require that an rdf processing system be able to
> distinguish between hard (extensional) and inferred 
> (intensional) facts. So
> you could ask for the type of #abanana and get the result 
> #banana by just
> querying hard facts; or query all facts (hard and inferred) 
> to find that
> #abanana is both a #banana and a #fruit. Or you build some 
> rules that can
> infer a #mostSpecificType property from all facts (though 
> probably only in a
> closed-world sense).

But that's the point. The fact you speak of (I believe) are
the explicit declarations of type for each literal value.

The issue here is not how to define data type
hierarchies and relations between data types, or to infer
knowledge about those relations, etc. RDFS already
provides for that. The issue is how to associate data types
with literals in a manner that results in a consistent
interpretation by arbitrary RDFS engines.

At present, interpretation is very "roll your own" and thus
there is a problem with portable statements about data type
membership of literals. I may generate all the qualified
anonymous nodes I like, but will they recieve consistent
intepretation in all contexts? And since it is a common view
(which I also hold) that qualified anonymous nodes are a
very "obese" form of representation, both in serializations
and in the graph, adopting a more compact packaging of data
types to literal values is of interest -- hence the concept
of URVs and the exploration of a consistent, standardized
interpretation for them (not just by RDF but by all URI aware
applications).

Cheers,

Patrick

--
Patrick Stickler                      Phone:  +358 3 356 0209
Senior Research Scientist             Mobile: +358 50 483 9453
Nokia Research Center                 Fax:    +358 7180 35409
Visiokatu 1, 33720 Tampere, Finland   Email:  patrick.stickler@nokia.com
  
Received on Friday, 19 October 2001 13:08:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT