W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > November 2002

Re: rdfs:Datatype question

From: Graham Klyne <Graham.Klyne@MIMEsweeper.com>
Date: Sun, 03 Nov 2002 10:35:50 +0000
Message-Id: <5.1.0.14.2.20021103103119.03e90590@127.0.0.1>
To: Brian McBride <bwm@hplb.hpl.hp.com>
Cc: "Patrick Stickler" <patrick.stickler@nokia.com>, <w3c-rdfcore-wg@w3.org>, "ext pat hayes" <phayes@ai.uwf.edu>

I think you're all correct, in different ways.  I think that, as Pat says, 
there's no fundamental need for rdfs:Datatype in the core language.  But 
there are applications for which it might be a useful piece of additional 
vocabulary.

I think that trying to define *every* piece of possibly useful vocabulary 
would be a slippery slope for this WG.  SO the question becomes:  what is 
lost by NOT having rdfs:Datatype in the core RDF(S) specs?

My own thoughts at this time are that someone might want to draft a NOTE 
about using datatypes as interpretation properties (which as far as I can 
tell is neither sanctioned nor forbidden by our work so far), and I think 
such a note would be a natural place to introduce a term with the intended 
meaning of rdfs:Datatype.

#g
--

At 11:41 AM 11/2/02 +0000, Brian McBride wrote:

>At 13:08 02/11/2002 +0200, Patrick Stickler wrote:
>
>>IMO, we need rdfs:Datatype to define the set of classes which
>>have the required characteristics for RDF datatyping, namely
>>a lexical space, a value space, and an N:1 mapping from the
>>lexical value space where N > 0.
>>
>>The term rdfs:Datatype is a means to give a name to the set
>>of RDF Classes which exhibit those characteristics.
>
>That is a good point, which I translate as: the model theory may say 
>nothing about the meaning of rdfs:Datatype, but would it be useful to 
>applications, e.g.  for example, knowing that something is a datatype 
>could trigger an app to go to its datatype implementation registry and 
>look for an implementation.  I'm not entirely convinced by that 
>example.  Maybe Patrick has one.
>
>Intuitively, it would seem a bit strange to have a concept like the class 
>of datatypes and not have a name for it.
>
>Brian

-------------------
Graham Klyne
<GK@NineByNine.org>
Received on Sunday, 3 November 2002 06:01:47 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:53:56 EDT