RE: How do RDF and Formal Logic fit together?

> From: Seth Russell [mailto:seth@robustai.net]
> From: "Pat Hayes" <phayes@ai.uwf.edu>
> 
> > >>  No way to indicate scope or variable bindings, chiefly.
> > >
> > >Well actually there is if you will allow that a scope of a 
> variable can
> be
> > >specified by a set of statements and that statements 
> themselves have
> > >identity.
> >
> > OK, no way to indicate a set of statements. Same thing.
> 
> Well actually there are two ways to do that.
> 
> 1) Point out the extension of the set with labeled directed 
> arcs by giving
> each statement an identifier:
> quad is (stid, subject, predicate, object)
> http://robustai.net/mentography/contexts.gif
> 
> 2) Include the name of the set in the label of the arc:
> quad is (subject, context+predicate, object)
> http://robustai.net/mentography/reificationContext.gif

Unfortunately, neither of these are RDF; one extends RDF whilst the other
applies a specific interpretation to something within RDF as it exists now.
That's fine if we do any of:

1) Extend RDF to include one of these mechanisms (thereby changing the
meaning of 'RDF' in a way that may or may not be backwardly compatible with
RDF 1.0);

2) Throw away RDF and make something else, maybe even called RDF (thereby
changing the meaning of 'RDF' in a way that is probably incompatible with
RDF 1.0);

3) Declare that RDF will never be used and that everyone should standardise
on a higher layer that provides such an extension (thereby removing RDF from
any real use other than as an encoding that can be done more effectively in
other ways).

It's not so fine if the following happens:

4) Some people use vanilla RDF, some people use an extension, and the two
argue about what constitutes 'RDF' and whether a 'RDF parser' can extract
anything useful from the other group's data (thereby muddying what
constitutes 'RDF').

[Caution: rant ahead]

I think we are in danger of (4) happening.  That's not a good idea in a
standards forum.  Unfortunately, it appears that the RDF-Core charter is not
broad enough for them to be able to consider (1) or (2); worse, it seems
that a number of large players are attempting to rush through RDF
standardisation so that each can use it to create incoherent and
incompatible graphs that none of the others can read, in the name of
standards.  This means that (3) may not happen and those players are trying
to limit the possibilities of (1) and (2) as well.

I haven't seen anything like it since HTML 3.2 --- but then I've kept my
nose out of the rest of the standards groups.  Maybe standards rally are
like sausages --- if you think you like 'em, don't watch 'em being made.

[Rant ends]

		- Peter

Received on Wednesday, 24 October 2001 16:59:45 UTC