Statements in RDF model

In discussions elsewhere, the question of whether or not the same statement 
(i.e. same subject/predicate/object triple) can appear more than once in an 
RDF model has been raised.  The expert opinion is that the use of 'set' in 
the following phrase from the formal model part of the M&S spec is quite 
deliberate, hence the answer is "no".

   There is a set called statements, each member of which is a
   triple of the form {pred, subj, obj}

This being the case, I would recommend that future versions of the RDF M&S 
specification be more explicit about this fact.

#g
--


P.S.  I'll also note that I have mild reservations about this aspect of the 
specification.

My view is that multiple occurrences of a statement should be recognized as 
distinct (even if, in many contexts, they may convey equivalent 
meanings).  I have two reasons for this:

1.  I think the RDF specification aims to say as little as possible about 
what any particular RDF expression actually means.  The idea that multiple 
properties of a single subject are to be interpreted as a conjunction seems 
to be no stronger than a commonly useful convention.  Is it conceivable 
that the same statement said twice is in some way different from being said 
just once?

2.  The reification model clearly allows separate statements with identical 
subject/property/object to be described.  I think that there should be a 
clear 1:1 correspondence between what can be described by reification, and 
actual RDF models (modulo consideration of reified statement resource 
identifiers as existentially quantified variables).

In <http://www.w3.org/DesignIssues/Toolbox.html>, TimBL has said:

   As RDF has little power at its basic level, anything new has to be
   introduced by reification - by describing it in RDF.

To me, this suggests that anything that can be described by reification 
should be allowed as a possibility in the RDF graph thus 
described.  Including multiple instances of the same 
subject/predicate/object triple.


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

Received on Thursday, 11 May 2000 12:56:37 UTC