RE: Question about rdf:value versus element content data graph re pres entation

Right. Duh. Serves me right for trying to work on the 
weekend, and with a fever on top of that... ;-)

In fact, global attribute bearing property elements are
really just a short hand for anonymous node objects. I.e.

<myproperty rdf:value="somevalue"/> 

is rather a short hand for

<myproperty>
   <rdf:Description>
      <rdf:value>somevalue</rdf:value>
   </rdf:Description>
</myproperty>

instead of being a shorthand for the content data literal
form.

Got it. (Don't know why I didn't before...)

Though it does make me start wondering about how one
defines or applies property constraints on values that
can be either literals or anonymous nodes acting as the
root of (possibly complex) subgraphs; and what
the relationship may be between graphs derived
from syndication of different forms -- where one 
literal value is defined as content data and the other
as rdf:value or some other anonymous node serialization;
e.g.

Source 1: <myproperty>somevalue</myproperty> 
Source 2: <myproperty rdf:value="somevalue"/>

One would like to think that the above two statements
represent equivalent knowledge, but since they end up
as different subgraphs hanging off the property node,
how does one distill them to achieve non-redundancy?

And if e.g. a DAML constraint says that 'myproperty'
values must be literals, does the second value satisfy
that constraint, since it is, in fact, an anonymous node?

Is it possible to write a schema constraint (RDFS or DAML)
that requires that only one or the other serialization be
used?

Just trying to get a better grip on aspects of genericity,
consistency, and portability of RDF encoded knowledge across 
arbitrary SW systems and agents...

Thanks,

Patrick

--
Patrick Stickler                      Phone:  +358 3 356 0209
Senior Research Scientist             Mobile: +358 50 483 9453
Software Technology Laboratory        Fax:    +358 7180 35409
Nokia Research Center                 Video:  +358 3 356 0209 / 4227
Visiokatu 1, 33720 Tampere, Finland   Email:  patrick.stickler@nokia.com
 

> -----Original Message-----
> From: ext Dan Brickley [mailto:danbri@w3.org]
> Sent: 08 September, 2001 13:06
> To: Stickler Patrick (NRC/Tampere)
> Cc: www-rdf-interest@w3.org
> Subject: Re: Question about rdf:value versus element content 
> data graph
> repres entation
> 
> 
> 
> Hi Patrick,
> 
> On Sat, 8 Sep 2001 Patrick.Stickler@nokia.com wrote:
> 
> > I have a question that I'm hoping someone with more insight to
> > the RDF spec than I have can provide.
> >
> > My (present and possibly incorrect) understanding of the difference
> > between the RDF XML '<myproperty>somevalue</myproperty>' and the
> > condensed variant '<myproperty rdf:value="somevalue"/>' is 
> that these
> > are only syntactic variants defining precisely the same knowledge
> 
> I don't believe this to be the case, although I agree the old 
> Model and
> Syntax spec is a little vague about the role/purpose of rdf:value. Is
> there a particular section of the spec that gave you this impression?
> 
> A number of apps that I've seen do use rdf:value and an intermediate
> resource when they want to associate, for example, data 
> typing or language
> tagging meta-information with literal values in a way that 
> shows up in the
> graph.
> 
> Dan
> 

Received on Saturday, 8 September 2001 06:45:10 UTC