Re: Trust, Context, Justification and Quintuples

On Fri, Dec 19, 2003 at 10:06:36PM +0100, Jeremy Carroll wrote:
> RDF Concepts is much better :)

Yeah, Concepts is cool, too, but it is mainly informative, and refers to
the RDF Semantics as the normative document for the parts that interest
me most :)

> OK we look at RDF Semantics 3.3.1
[...]
> basically Pat is listing all the things you think might be true and
> pointing out that they do not necessarily hold.

But I don't need it to _necessarily_ hold, _may_ hold is good enough :)

In the paragraph between the two you quoted there is a passage:

http://www.w3.org/TR/rdf-mt/#Reif:
> This particular interpretation of reification was chosen on the basis
> of use cases where properties such as dates of composition or
> provenance information have been applied to the reified triple, which
> are meaningful only when thought of as referring to a particular
> instance or token of a triple.

It explicitly states applicability of reification to description of
"dates of composition and provenance information". Seems like your
requirements to provenance information differ from Pat's ;-)

> The clearest examples are ones in which we have owl:sameAs to say that two 
> urirefs have the same denotation.

Back from student times I remember that when professor says "obvious",
usually it is not... ;-)

> Here we go ..
> 
> eg:a owl:sameAs eg:b .
> eg:r rdf:type rdf:Statement
> eg:r rdf:subject eg:a .
> eg:r rdf:predicate owl:sameAs .
> eg:r rdf:object eg:b .
> 
> SinceI(ag:a) = I(eg:b) we have that this graph means exactly the same
> as the following graph
> 
> eg:a owl:sameAs eg:b .
> eg:r rdf:type rdf:Statement
> eg:r rdf:subject eg:a .
> eg:r rdf:predicate owl:sameAs .
> eg:r rdf:object eg:a .
> 
> i.e. if eg:a owl:sameAs eg:b is true then its reification is also the
> reification of eg:a owl:sameAs eg:a (and eg:b owl:sameAs eg:a, and
> eg:b owl:sameAs eg: b). 

Doesn't it contradict with "there are no entailment relationships which
hold between a triple and a reification of it"?

> The classic example is "I saw Superman fly" and "I saw Clark Kent fly"
> are two different sentences even if we know that Superman = Clark
> Kent. RDF Semantics does not allow you to capture this using
> reification.

I agree with Sandro's solution to this example.

> For trust mechanism, provenance etc. this really is not what is wanted
> - a true quoting mechanism leaves the urirefs unchanged and does not
> interpret them. (Although I think it is OK to rename bnodes - not
> thought that one through properly yet) 
[...]
> Thus any quad mechanism, or the named graph stuff, associates the quad
> ID or the graph name, not with the interpretation of the s,p,o but
> with the actual nodes s, p, o - quoting them, rather than interpreting
> them.

I still don't see why is it necessary, and why you can't use something
like log:uri to retain urirefs?

-- 
Dmitry Borodaenko

Received on Thursday, 5 February 2004 11:33:16 UTC