Re: Hidden triples and self-description

Mark Baker wrote:

>I think that's different.  All parties may know RDF, but those that know
>the rules of entailment will extract additional triples from the same
>message.  Hence my comment about media types; if it is the intent of the
>sender to communicate this additional triple, then using the RDF media
>type isn't sufficient.
>  
>
I don't see this as any different than for any application applying any 
other semantics to any other syntax. For example, any 'ol XML. The 
sender and receiver may assume a _meaning_ for the XML message but that 
is not defined by the XML _itself_. The XML transmits syntax _onto 
which_ an application applies semantics.

Those 'additional triples' are _not_ part of the syntax of the message, 
rather the model theory licenses these as entailments along the lines of 
"given this message, I assume these entailments"

>  
>
>>Indeed the RDF and OWL 
>>model theories are published in well known places, so isn't it sort of 
>>obvious that if you are going to use the RDF or OWL model theories to 
>>derive entailments, that you'd have to know the rules of entailment?
>>    
>>
>
>I don't know.  But I see that there's some disagreement over this in
>the community, e.g. (at least this appears to be talking about the
>same thing)
>
>http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/0090.html
>  
>
Not at all. An RDF parser emits the triples given a particular piece of 
RDF/XML. What an inference engine might infer from these triples is 
licensed by the RDF MT entailments. The RDF MT _does not_ in any way 
shape or form change what triples are actually in the RDF/XML. Indeed 
the RDF MT as well as the OWL MTs operate not at the level of RDF/XML 
rather on the (parsed) triples otherwise known as the _graph_.

Jonathan

Received on Monday, 15 September 2003 10:49:59 UTC