RDF model revisited, or: How to make the most out of Reifications and Containers

Please, allow to point readers of this list that are interested in the
RDF model, reification, or containers to a short discussion paper we
wrote recently. Any comment is welcome (conen@gmx.de,
klapsing@wi-inf.uni-essen.de). Below, you'll find the abstract and two
additional remarks.


Thank you for your kind consideration,
      Wolfram and Reinhold


TITLE: RDF model revisited, or: How to make the most out of Reifications
and Containers

AUTHORS: Wolfram Conen and Reinhold Klapsing

ABSTRACT: The basic result presented is the following: with two
(hopefully) reasonable assumptions about the intentions behind the RDF
model, it can be shown that the RDF model and a model allowing for
nested triple and lists of resources and triples, can be mapped to one
another (in both directions). This allows to establish a close link
between the RDF model and "extended" models recently suggested (Slim
RDF, XRDF). Furthermore, it can be used to define the notion of
essential models that allows to get rid of unneccessary triples in
reifications and containers while still capturing the underlying
intentions of the original RDF model.  This, in turn, may help to
clarify some problems related to interpreting the roles of reification
and Bag, Alt and Seq in the RDF model.

LINK: http://nestroy.wi-inf.uni-essen.de/rdf/RDFmodel_revisited_v10.pdf


Two additional remarks:

* We do not propose to change the RDF model (I hope that you like to
read this, Brian ;) -- instead, we "interpret" the RDF model in a
certain "upwards compatible" way, allowing to "identify" reification
with nesting and containers with lists.

* Reification and containers are treated as instruments that can be used
for automatic structural transformations. (I hope that using the word
"structure" here is not too far fetched - Graham? Pat?). That means,
that the more "sophisticated" uses of the "reification construct" in RDF
would have to be defined (precisely/formally) on a semantic/logic level
(on the "schema" level, to use RDF speak). The same is true (more or
less) for containers (especially for the meaning of a container's type).
In this sense, the discussion paper is an approach to "clean" RDF (not
RDFS) from hidden and ambigous semantic implications and to interpret it
as a way to express nested triples/lists (and that's it).

Received on Thursday, 31 May 2001 11:07:58 UTC