Re: DAML+RDFS: potentials for simplifications?

On Thu, 29 Nov 2001 Joachim.Peer@unisg.ch wrote:

> ==> implications of RDF for flexible (i.e. upgrade-able)
> software agents: the core question for me is, if RDF
> enables some kind of generic extension framework for
> software agents.
>
> it would be cool to be able to "upgrade" software agents
> simply be importing an extension module which (in your
> case) could  contain:
> - a number of RDF triples (representing axioms of a higher
> level ontology schema like daml+oil)
> - a number of RQL queries, that provide some kind of
> inference patterns.
>
> is this the vision you are working on?

To be honest I am mostly working on getting the thing to
properly grok RDF Schema :)

However, in thinking about upgrading to DAML+OIL support I
had mostly thought of hard-coding its semantics in some
fashion, and not adding it flexibly in the way you suggest.

For me the benefit of having DAML+OIL expressed in RDF is in
the fact that I can add support for DAML+OIL in a gradual
fashion, in the meantime having a tool that can at least
partially understand it. Sesame already understands RDF
Schema semantics, and I can (hopefully) gradually extend its
model to incorporate DAML+OIL primitives.

I am also unsure whether this idea of flexible addition of
expressivity is feasible at all. The problem to me seems to
be that you need a starting point that is itself at least as
expressive as the language that you are trying to "learn",
which kind of defeats the purpose of the undertaking.

RDF does not provide enough expressiveness to represent
axioms. How would you encode the DAML+OIL language axioms in
RDF, in such a way that a software agent that _doesn't_
understand DAML+OIL can interpret them?

Or am I completely missing the point here?

> how likely is this vision realizable?

I am a bit pessimistic, but maybe I am just not in tune with
the idea yet.

> or will/should DAML+OIL be the conceptual basic framework
> for such an effort (which in turn, would make RDF
> serialization less relevant and could be generated ex
> post, for compatibility reasons)

I think the main benefit of the DAML+OIL / RDF connection is
not so much in the XML serialization syntax, which we all
agree is rather eh... convoluted.

The important point is the connection between the _models_
(graphs/triples). DAML+OIL is only more expressive than RDF
because it gives an agreed-upon meaning for a couple of node
and edge labels. But I do not see how you can convey this
meaning to a software agent without hard-coding it.

But the idea is certainly interesting and I would happily be
tought that it is - in principle - possible, it would
certainly open cool possibilities for my own project :)


Regards,

Jeen
-- 
Jeen Broekstra                                     Vrije Universiteit
jbroeks@cs.vu.nl              Dept. of Mathematics & Computer Science
                                 de Boelelaan 1081, 1081 HV Amsterdam
                                                      The Netherlands

Received on Thursday, 29 November 2001 10:10:02 UTC