Re: discussion strawman: RDF Data Model Summary

Good work!  The rdfmodel document proves that the basic tenet of RDF Model
is sound and quite straightforward, once the complexity the syntax brings
has been removed.  I feel confident that such an overview would be far
easier to comprehend by the newcomer.  Hopefully any alternative syntax we
can devise will be equally straightforward.

I have a few suggestions for the second pass.  It mentions distributive
referents in the context of containers, but does not go on to describe this
aspect.  Also, there is no mention of anonymous resources and their two
primary uses - qualified property values and non-binary relations.  It would
be good if you could include these important and powerful aspects of RDF
Model in the document.

How about a second revision, please :)

It already highlights several issues that I feel would need resolving (all
of them mentioned in the group before, I think):

1) Section 3 states that container resources must be declared to be an
instance of one of the container object types (Bag, Seq or Alt).  I would
prefer this to be that container resources must be declared to be an
instance of (a subclass of) rdfs:Container.  There are three subclasses
provided (in the absence of any suitable schema) as standard with RDF: Bag,
Seq or Alt.

2) Section 5 states that Properties are (a subset of) Resources.  Is this
really the case?  Can you really use a property wherever a Resource is
expected?  How about as a statement subject?  If they are, there is no way
to directly attribute a URI or ID to them with the current syntax (an ID
would identify the reification of that statement).  If they are not, it
would clear up the RDFSchema issue of where subClassOf ends and
subPropertyOf begins - you could have 'subClassOf' and 'subPropertyOf' for
the seperate Resource and Property type hierarchies respectively.

3) Section 5 should state that statements are also (a subset of) Resources -
I believe that rdf:Statement is a subclass of rdfs:Resource in RDFSchema.
Hence you should also be able to attribute URIs to their reified
representation as well.

4) I get the impression that collection member ordinals (_1, _2, etc) were
designed this way to allow them to be abbreviations of the <rdf:li>
properties as attributes in the current RDF syntax.  Several significant
drawbacks to this have already been raised in this group and in
www-rdf-comments.  I would argue that it is not pure model stuff (and I
would not want to restrict any future syntax to following this).  Hence I
believe it doesn't belong in the rdfmodel doc.

What do you think?

Received on Saturday, 9 September 2000 19:18:26 UTC