Re: making statements about a graph

Max Froumentin wrote:
> Steve Harris <S.W.Harris@ecs.soton.ac.uk> wrote:
> 
> 
>>The RDF provided mechaism for reification is pretty messed up IMHO.
> 
> 
> This reminds me that I never saw a clear statement explaining why it
> was messed up. I want to use it for voice recognition, to express
> confidence values for assertions:
> 
> <utterance id="u1">chat</utterance>
> 
> <rdf:RDF>
>   <rdf:Resource rdf:nodeID="r1" rdf:about="#u1" dc:language="EN"/>
>   <rdf:Resource rdf:nodeID="r2" rdf:about="#u1" dc:language="FR"/>
> 
>   <rdf:Resource rdf:about="#r1" foo:confidence="65%"/>
>   <rdf:Resource rdf:about="#r2" foo:confidence="35%"/>
> </rdf:RDF>
> 
> Anything broken with that? 

Aside from the validation errors and undefined element rdf:Resource, you 
mean?  Use the rdf validator

http://www.w3.org/RDF/Validator/ARPServlet

Even if it were legal rdf, it would not do reification in the rdf sense, 
and would probably not indicate what you want.  That is because the 
element named "utterance" is not something known to rdf.  It is not a 
"resource" in the rdf sense, and it certainly is not an rdf statement.

There is nothing to stop you from writing an application that knows what 
to do with your construction, but it will not be understood by standard 
rdf processors.

In rdf, a reified Statement is not asserted, that is, there is no actual 
statement with the specified subject, object, and predicate in the 
graph.  There is no way to connect a specific Statement (a reified 
triple) to a specific triple in the graph.

Cheers,

Tom P

Received on Thursday, 6 November 2003 09:32:03 UTC