W3C home > Mailing lists > Public > www-rdf-interest@w3.org > June 2003

Re: RDF's binary nature

From: Francesco CannistrÓ <fracan@inwind.it>
Date: Thu, 19 Jun 2003 18:55:32 +0200
Message-ID: <035601c33683$99d8e510$38971d97@Matrix>
To: <www-rdf-interest@w3.org>, "Tom Reilly" <treilly@macromedia.com>

It is valid RDF. A property can be asserted multiple times (with possibly
different values) for the same resource. Whether this is or is not meanigful
it depends on the the semantic  of the property.
Consider the "isMotherOf" property. Maybe you want to assert that Monica has
2 childs (Maria and Tom):

<rdf:Description rdf:about="#Monica">
   <isMotherOf rdf:resource="#Maria "/>
   <isMotherOf rdf:resource="#Tom"/>
</rdf:Description>

But if you consider a property like "hasMother", then if you write:

<rdf:Description rdf:about="#Tom">
   <isMotherOf rdf:resource="#Elena"/>
   <isMotherOf rdf:resource="#Monica"/>
</rdf:Description>

these assertion are clearly meaningless.

RDF(S) actually does not care to provide means to assert the cardinality of
properties: whatever assertion your description contains it is valid RDF
since it respects syntax and schema constraints.  You can restrict the
cardinality of a property only with OWL

cheers

----- Original Message -----
From: "Tom Reilly" <treilly@macromedia.com>
To: <www-rdf-interest@w3.org>
Sent: Thursday, June 19, 2003 6:00 PM
Subject: RDF's binary nature


>
>
> Since I first started reading about RDF I assumed its binary
> nature implied that you couldn't have a resource with more
> than one value for the same property, ie that this wasn't legal:
>
> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>          xmlns="http://www.foo.org/bar#">
>   <foo>
>     <bar>baz</bar>
>     <bar>boo</bar>
>   </foo>
> </rdf:RDF>
>
> But apparently it is (it validates with the RDF Validation Service
> and I've encountered real world examples of the same basic structure)
> and when I think about it there's no real basis for such an assumption.
>
> I was hoping someone could verify that this is in fact valid and possibly
> explain why this is legal when we have bags.  The fact that there are
> no examples of this in any of the RDF specs I've read (that I can remember
> at least) could be seen as a little misleading.
>
> I'm trying to come up with a simple RDF API and the API could be much
> simpler if this wasn't the case.  I also realize that this structure
> is very common in other XML dialects and disallowing it would create a
> barrier to converting them to RDF.
>
> Also if I have:
>
>     <bar>baz</bar>
>     <bar>baz</bar>
>
> instead I'm making two distinct but otherwise equivalent statements right?
>
> Thanks in advance...
>
>
Received on Thursday, 19 June 2003 12:55:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:59 GMT