Re: abstract model and reification

At 03:07 PM 9/18/00 -0700, Seth Russell wrote:
>Graham Klyne wrote:
>
> >    [A] --P-------------> [B]
> >    [S] --rdf:type------> [rdf:Statement]
> >    [S] --rdf:property--> [P]
> >
> > ?  Here, S may or may not be a reification of [A] --P--> [B].  But from its
> > type, it clearly represents _some_ reified statement.
>
>Seems to me that would represent the reification of all arcs labeled with "P".

Reasonable.  Or, more precisely, an incomplete reification of all arcs 
labelled "P".

Adding "[S] --rdf:subject--> [foo]" found further constrain that 
interpretation.


>I'm having troubles following this discussion on reification because I haven't
>found a write up on the basic motivation for reification in the RDF 
>model.   In
>particular im trying to find answers to the following questions:
>1) Where is it necessary to reify and where not?

To make statements about statements.  I say "[the sheep] is [black]", but 
you might say "[the sheep] is [white]".  Reification allows us to model 
these statements *using the very same graph model that is used to assert them*.

>2) If we want to say something about a statement that is asserted in some
>context, why cant we just designate it's id as an object (see [1])?

In a sense, that's exactly what reification does.  The object thus 
designated is called a "resource" and is said to have properties rdf:type, 
rdf:subject, rdf:object, rdf:property as specified.

>3) If all RDF statements (reified or not) are contained in some context, then
>why isn't that context alone sufficient to disambiguify any statement about
>another statement?

Well, I am working on some ideas in that space.  The base RDF graph has no 
formal concept of "context", though there is an idea of "containing 
document" that I find rather unsatisfying.

>4) Where are there some concrete examples to motivate all of the above?

The RDF spec has a few.

#g

------------
Graham Klyne
(GK@ACM.ORG)

Received on Tuesday, 19 September 2000 09:31:56 UTC