W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2000

Re: A DTD for the abstract RDF model

From: Jonas Liljegren <jonas@liljegren.org>
Date: 09 Oct 2000 17:10:10 +0200
To: Pierre-Antoine CHAMPIN <champin@bat710.univ-lyon1.fr>
Cc: Graham Klyne <GK@Dial.pipex.com>, Tom Van Eetvelde <tom.van_eetvelde@alcatel.be>, www-rdf-interest@w3.org
Message-ID: <87u2am2gq5.fsf@jonas.rit.se>
Pierre-Antoine CHAMPIN <champin@bat710.univ-lyon1.fr> writes:

> I guess we could get rid of any llteral by writing :
> 
> <Description>
>   <value resource="data:98">
>   <type resource="http://www.datatypes.org/useful_types#Integer"/>
> </Description>

I think we should make a diffrence between the value used to describe
a property and the property itself.

Let's say we have these three URI's:

 #A : represents the a spcific person
 #B : represents the name of that person
 name : the predicate

 We have the statement:
 #A --name--> #B

The value of #B can be 'Axel'.  But #B can't be used as a literal for
other persons names.  If Axel changes his name to 'Sture', the
resource would still be the same.


Maby this example will be more clear if you consider the properties
'age', 'temperature', 'height', etc.  All these properties consists of
simple values.  But they can't be used by other properties, even if
they happen to have the same value.  They are 'dynamic' literals.


I would like all literals to be seen as dynamic.  Other things, like
places, coulers, etc, that can be used in many properties, will
themself be specific resources.

Words that has the same spelling but diffrent meanings will have
diffrent resources.


But yes.  We should also have a standard URI for each number etc.
Even if they should be used with caution.





> Jonas Liljegren wrote:
> > 
> > Graham Klyne <GK@Dial.pipex.com> writes:
> > 
> > > I think I can accept the broad thrust, but maybe have problems with
> > > a few details.
> > >
> > > If I understand correctly, you replace a literal with a resource
> > > with properties, as in:
> > >
> > >    <Description>
> > >      <value>98</value>
> > >      <type resource="http://www.datatypes.org/useful_types#Integer"/>
> > >    </Description>
> > >
> > > so that the 'value' attribute has the literal value as an object, the 'type'
> > > indicates a subtype of literal.  Attributes of the property referencing the
> > > literal, such as xml:lang, can be represented as properties on the literal
> > > resource.
> > 
> > Yes.  And this is based on the RDF M&S and RDF Schema.  The M&S have a
> > section talking about the value property, used to specify the unit for
> > a value, and other things.  The RDF Schema defines the Literal class.
> > 
> > > Given the above, what (if anything) is to prevent:
> > >
> > >    <Description>
> > >      <value>
> > >        <Description>
> > >          <value>98</value>
> > >          <type resource="http://www.datatypes.org/useful_types#Integer"/>
> > >        </Description>
> > >      </value>
> > >      <type resource="http://www.datatypes.org/useful_types#Integer"/>
> > >    </Description>
> > >
> > > ?  i.e. repeated substitution of the literal with its interpretation as a
> > > resource.  Does this matter?  Maybe not.
> > 
> > Exactly. But that should only be a syntax thing.  We can now simply
> > remove the literal from the DLG. ... But the thing here is that that's
> > optional.  Everything will work the same way either way.
> > 
> > > When you wrote:
> > >
> > > >Age as a resource:
> > > >
> > > ><Description about="http://give.me/a.number">
> > > >     <type resource="http://www.datatypes.org/useful_types#Integer"/>
> > > ></Description>
> > >
> > >
> > > I'm not sure how you mean the value "98" to be conveyed.
> > 
> > The number will be conveyd by the implementing software, outside the
> > realm of RDF.  Either by accessing some database or internal
> > representation to get the value associated with the resource.  Or maby
> > by accessing the page http://give.me/a.number and displaying the
> > retrieved result.
> > 
> > The method of accessing the actual value will be dependent on the
> > implementing software.  Let's say that the resouce
> > http://give.me/mp3.file represents a mp3 music file.  We could say
> > that the URI http://give.me/mp3.file#title will represent the title of
> > the track, stored inside it.  A mp3-aware RDF program could retrieve
> > the actual title by parsing the mp3 file.
> > 
> > For more about all sorts of RDF things. Check out the mail archive for
> > the Wraf project:
> >         http://www.uxn.nu/wraf/
> >         http://www.uxn.nu/cgi-bin/mailman/listinfo/rdf
> > 
> > --
> > / Jonas  -  http://jonas.liljegren.org/myself/en/index.html
> 
> -- 
> --- Quid quid Latine dictum sit, altum viditur
>     Whatever is said in Latin sounds important.

-- 
/ Jonas  -  http://jonas.liljegren.org/myself/en/index.html
Received on Monday, 9 October 2000 11:09:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT