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

RE: RDF speficiations (was RE: Cutting the Patrician datatype kno t)

From: <Patrick.Stickler@nokia.com>
Date: Sat, 8 Dec 2001 00:51:41 +0200
Message-ID: <A03E60B17132A84F9B4BB5EEDE57957B160B1E@trebe006.NOE.Nokia.com>
To: tarod@softhome.net, www-rdf-interest@w3.org


> -----Original Message-----
> From: ext tarod@softhome.net [mailto:tarod@softhome.net]
> Sent: 04 December, 2001 04:11
> To: www-rdf-interest@w3.org
> Subject: Re: RDF speficiations (was RE: Cutting the Patrician datatype
> knot)
> 
> 
> 
> 
>    Sorry, I'm not following this thread but I guess you are 
> talking about
> giving meaning to literals. I mean use data types for 
> literals, so, instead
> of having...
> 
>    John is 23
> 
>    I would have
> 
>    John is age1234
>    age1234 type integer
>    age1234 value 23
>  
>    Is that right?

Yes. This follows the DAML idiom. Though you'd probably want
the property to be 'age' rather than 'is'. E.g.

     John age 1234
     1234 type integer
     1234 value 23

You could also infer the typing globally in a schema
by using rdfs:range, where you'd say
   
     John age "23"

and define that age has an rdfs:range of integer.

>    That suposes that we will have an instance of RDF with the 
> triple of
> sentences for each literal we had in the original one. When you have a
> really big model this is not posible, 

This is a legitimate practical concern, and one that URVs
are intended to address (without loosing the explicitness
of localized typing).

C.f. http://www-nrc.nokia.com/sw/X_Values_URI.pdf

> I think that this is 
> ok, if you need
> it but using a new extension, name it DataTypedRDF DTRDF or 
> something like
> this but if you want to use the model without typing, why 
> not???? It's a
> model that doesn't know anything about semantics 010 is 
> different that 10,
> two literals 10 and 10 are diferent, you must compare them in your
> application, maybe you will never need this comparation, so, 
> why add them
> to the RDF Core? RDF Core must be as easy as possible, RDF 
> Schema should
> be, too, an each extension over a basic model will be more 
> complex, but you
> will use it, only if you need it.


I fully agree with you that RDF should only capture the minimal
knowledge that enables an application to make these comparisons
and interpretations. Data typing is very important, and RDF needs
to provide a clear methodology that works well, but IMO it should
be as generic and flexible as possible.

Cheers,

Patrick
Received on Friday, 7 December 2001 17:51:50 GMT

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