RE: Predicates and Arcs vs Triples RE: use/mention and reificat ion: rdf:predicate/subject/object

> From: Geoff Chappell [mailto:geoff@sover.net]
[...]
> My understanding of the rdf roadmap is that RDF
> will serve as a triple-based data model for storage of facts

The problem that's slowly being exposed here seems to be that while RDF is
capable of storing data (with no interpretation), it isn't capable of
storing facts unless RDF has a firm semantic basis, because a fact is data
as interpreted by some processor or logic system.

If you put that firm semantic basis under pure RDF, its expressive power is
either very limited (without reification) or slightly less limited but
rather peculiar (with reification).  For example, neither approach seems to
allow negation, as it does not seem to be possible to build the notion of
negation into the current version of RDF.

If, instead, you want to allow higher layers to make use of RDF, it seems
that some encoding is required to distinguish:

(a) an RDF fact that can be interpreted using only RDF semantics, and is
interpreted identically by a pure RDF engine or by any higher layer that
uses the same semantics for interpreting RDF facts, versus

(b) a data structure that happens to be expressed in RDF but can only be
interpreted by a higher layer, and may not be interpreted identically by
different higher layers.

> Geoff Chappell

		- Peter
--
Peter Crowther, VP Development, Network Inference Limited.
http://www.networkinference.com

Received on Friday, 1 June 2001 11:40:42 UTC