W3C home > Mailing lists > Public > www-rdf-comments@w3.org > January to March 2003

RDF Semantics: Interpretations and Modelling

From: Ossi Nykänen <onykane@butler.cc.tut.fi>
Date: Mon, 3 Feb 2003 16:44:09 +0200 (EET)
To: www-rdf-comments@w3.org
Message-ID: <Pine.GSO.4.44.0302031636240.27550-100000@butler.cc.tut.fi>

#1.) RDF semantics does not define interpretation for containers or
collections (etc.). Since RDF semantics treats URIs as parameters, this
means that RDF models containers can interpreted arbitrarily. This seems
to nullify the work and efforts spend on defining, e.g., the rdf:bag
container or rdfs:comment. (From the viewpoint of entailments, why bother
using bag if it's "just" a URIref.)

In particular, this seems to make RDF an "extension" of itself since RDF
syntax, RDF Schema, and RDF Concepts all go on defining and using more
vocabularies than the RDF Semantics defines (the idea of a model theory
does not exclude [error-prone] narrative descriptions of interpretations).

Perhaps defining some of these interpretations at least as an informative
appendix would be in order? (The bag and collection, please ;)

#2.) The role of URIrefs as parameters (in the context of entailment)
might be worth spelling out in the RDF Concepts as well. In particular,
RDF Concepts seems to treat blank nodes as anonymous building blocks to
construct complex statements using binary sentences. Semantics, however,
treats them as something that MUST be instantiated in the end. Knowing
this, most RDF designers would thus probably want to steer the set of
interpretations by using URIrefs instead of blank nodes? ("These nodes are
only intended to be a SLOTS in a structure, not e.g. particular
[different] entities whose names I don't care of.")

#3.) Since RDF operates with URIrefs, the universe of a model is in the
draft designed to operate with resources. This follows is the traditional
correspondence theory of truth -way of things.

However, for practical reasons the set of resources could as well be
replaces with an (other) set of URIrefs. This would allow speaking
directly about pieces of RDF models by using RDF itself. (And we would not
lose anything since the resources "themselves" are out of reach anyway.)

--Ossi
Received on Monday, 3 February 2003 09:44:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:31 GMT