Some comments, below. (01)
Thanks,
Leo (02)
-----Original Message-----
From: uos-convene-bounces@xxxxxxxxxxxxxxxx
[mailto:uos-convene-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Bill
Andersen
Sent: Saturday, March 04, 2006 10:53 AM
To: Upper Ontology Summit convention
Subject: Re: [uos-convene] KBR vs SQL (03)
On Mar 4, 2006, at 09:26 , Cassidy, Patrick J. wrote: (04)
> Mathew,
> Since the large ontology-based applications that will be useful
for
> major problems are likely to use relational databases (or their
> equivalent) as their back-end, what is the problem with saying that
> the
> capabilities of such systems are more advanced than the
> capabilities of
> database systems alone? (05)
Because it's hype that has not been substantiated by experience.
Hype is bad. If we're not careful we're going over-hype all of this
and create ourselves another "AI winter". (06)
Secondly, you have to be careful about what you mean by saying "more
advanced". Take Oracle's RDFS store for example. Many consider RDFS
to be "semantic technology" or even "ontology" (god help us!). It is
built as a straightforward extension of Oracle's relational
technology. Now, do you think RDFS is "more advanced" than relational? (07)
LEO: I would say RDF/S IS MORE EXPRESSIVE than the relational model. It
is more like Entity-Relation or Extended-ER, as a positive existential
subset of FOL (as Pat Hayes would say). I don't know what "more
advanced" means. If it means "more expressive", than it is more
advanced. Sured RDFS is a very simple ontology language, so what? It
does have a model-theoretic semantics behind it, which is more than
I've ever seen for ER or EER. (08)
LEO: To me, the advances in ontological engineering in recent years has
not been due to AI. Its been due to formal ontology, formal semantics,
and FOL-based technologies. The AI part is primarily the Description
Logic part (though of course the non-monotonic, defeasible, etc.,
logics thread through the late 80s/early 90s was a good thing), and
that is a bit problematic. Hype is always bad, but I see nothing
similar to the use of semantic technologies on the horizon to
prospectively (p r o s p e c t i v e l y) offer a qualitative leap (I
want to say "revolutionary", but will mind my rhetoric) over current
state of the practice in computing/information technology. What's the
alternatives? Genetic programming/evolutionary computing/complex
adaptive systems? Ok, I won't hold my breath for those technologies as
contributing across the board any time soon. (09)
> They are indeed not "competing" but they are
> also not "complementary"; the larger ontology-based systems will
build
> on and extend the database systems. (010)
What "larger ontology-based systems"? Which ones do you have in
mind? What precisely do these systems do? I'm getting the feeling
that you're talking about something very comprehensive like Cyc... (011)
LEO: No, what he is talking about is a semantic level that relational
databases don't currently have.
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files:
http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (012)
|