Re: DECIDED: untidy semantics

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext Jos De_Roo" <jos.deroo.jd@belgium.agfa.com>
To: "Brian McBride <bwm" <bwm@hplb.hpl.hp.com>
Cc: "w3c-rdfcore-wg" <w3c-rdfcore-wg@w3.org>
Sent: 24 September, 2002 01:04
Subject: Re: DECIDED: untidy semantics


> 
> 
> >At 13:17 21/09/2002 +0200, Jos De_Roo wrote:
> >
> >>[regretting not to have been there, but was impossible]
> >>
> >>taking http://ilrt.org/discovery/chatlogs/rdfcore/2002-09-20.html
> >>for the moment for granted
> >>
> >>[[[
> >>....
> >>[14:48:03] jjcscribe Vote: untidy 7, tidy 5.
> >>[14:48:11] jjcscribe DECIDED: untidy semantics.
> >>[14:48:22] jjcscribe DanC has outstanding dissent.
> >>]]]
> >>
> >>I'm recording outstanding dissent too !!!
> >
> >Reading what you say below, perhaps you are not dissenting.  The question
> >put was:
> >
> >[[
> >  where you have an rdf;descipriton element containing a property element
> >of <age>10</age> is that tidy or untidy.
> >]]
> >
> >You are objecting to
> >
> >   <a> <b> "foo" .
> >
> >being untidy but that
> >
> >   <a> <b> _:dt"foo" .
> >
> >is ok.
> 
> right, the latter is untidy
> thanks for the clarification Brian
> 
> >Then, the question as put, if it translated to:
> >
> >   <something> <age> _:dt"10" .
> >
> >would you be ok with that?
> 
> of course I would
> but then how would you write (in RDF/XML)
>    <something> <someproperty> "somestring" .
> which I consider as a very meaningful statement

Jos, which are you trying to express here:

1. That <something> has <someproperty> with the value denoted by "somestring".
2. That <something> has <someproperty> with the value of the string "somestring".

???

If 1, then the untidy, implicitly datatyped form is relevant:

   <rdf:Description rdf:about="#something">
      <someproperty>somestring</someproperty>
   </rdf:Description>

which equates to

   <something> <someproperty> _:x"somestring" .


If 2, then you should specify the datatype, locally or
globally, to explicitly assert the string datatype. E.g.

   <rdf:Description rdf:about="#something">
      <someproperty rdf:datatype="&xsd;string">somestring</someproperty>
   </rdf:Description>

which equates to

   <something> <someproperty> <http://...#string>"somestring" .

or alternately

   <rdf:Description rdf:about="#something">
      <someproperty>somestring</someproperty>
   </rdf:Description>
   <rdf:Description rdf:about="#someproperty">
      <rdfs:range rdf:resource="&xsd;string"/>
   </rdf:Description>

which equates to

   <something> <someproperty> _:y"somestring" .
   <someproperty> rdfs:range xsd:string .

which by the datatyping model theory, is semantically
equivalent to the locally typed case.

And note the compatability of the above two forms of expression. In
both cases, the literal node denotes some value, and whether that
value correlates to the lexical form in the label or some other
value is simply a matter of which datatype is specified.

> one can have
>    <something> <someproperty> _:somelabel"somestring" .
> when using an explicit rdf:nodeId attribute

Do you consider this a good or bad thing, that one could
use rdf:nodeID to refer to a particular occurrence of
an inline literal? It seems a useful and innocuous feature
to me.

Cheers,

Patrick

Received on Tuesday, 24 September 2002 05:40:51 UTC