Henry --

You wrote (below)...

Instead of trying to argue that there is a general problem with logic,
which seems to me to be a hopeless claim, it would be much better to
work on simple problems and find out how best to model them.

There's a strong logic foundation [1] to our system [2].  But that needs to be supplemented with executable (open vocabulary) English, which we do on top of the logic.

At [3] please find a number of examples of problems and their executable English models.

                                Cheers,  -- Adrian

[1] Backchain Iteration: Towards a Practical Inference Method that is Simple Enough to be Proved Terminating, Sound and Complete. Journal of Automated Reasoning, 11:1-22

[2] Internet Business Logic, online at www.reengineeringllc.com

[3]  http://www.reengineeringllc.com/demo_agents/

At 04:18 PM 2/9/2005 +0100, you wrote:

I would start from a completely opposite view point.
RDF is as powerful as first order logic or set theory. Which is incredibly powerful.

The problem is just that some good basic patterns on how to think
about various elements of reality still need to be established clearly
in the RDF community.

My Madonna example shows how one should think about objects temporally.
This is nothing very new. It is really an idea I got from reading David
Lewis very carefully 15 years ago.

The Foaf vocabulary has given some very good examples on how to identify people using inverse functional properties.

Instead of trying to argue that there is a general problem with logic,
which seems to me to be a hopeless claim, it would be much better to
work on simple problems and find out how best to model them.

Logic, like most of mathematics is based on such simple and indubitable
premisses, that trying to put them in doubt, especially for engineering
reasons, would be like seeing a bridge builder doubt that mathematics
was valid because his bridge fell apart.

Henry Story

On 9 Feb 2005, at 16:03, Adrian Walker wrote:
Frank, Henry, Murray , Bill --

Good discussion, (appended below) !

There was a nice exchange between Bill Kent and a leading academic around the time when the "universal relation"  (UR) was being proposed as a database model.  Bill pointed out that if the UR contained "person has horse" and "horse has birthday", then one could conclude "person has birthday" -- clearly wrong.

The story comes to mind because there seems to be a history of  IT folks proposing "semantic" models that unnecessarily leave out essential aspects of the real world.

The latest example of this is perhaps "data semantics" as practised in the XML and RDF/OWL communities.  Great as far as it goes, but only half the story.

The naming problem is difficult, but we seem to make it even more difficult by trying to solve it with amputated models of the real world.  Yet there are folks intent on doing just this for the US National Health Information Network and other very large projects.

Frank will be familiar with my argument that there is something practical that we can do about this right now.  This is to include executable, open vocabulary English as part of the modelling toolkit, and as part of the fielded systems.  In a sense, this adds real world application semantics to the current data semantics.

The e-Government presentation at www.reengineeringllc.com shows how to do this, and it's based on a practical system that is online at the same site.

How does this sound ?
                                                     Cheers,  -- Adrian

Adrian Walker
Reengineering LLC
PO Box 1412
CT 06011-1412 USA

Phone: USA 860 583 9677
Cell:    USA  860 830 2085
Fax:    USA  860 314 1029

At 09:26 AM 2/9/2005 -0500, you wrote:

It's available online.  Go to http://www.authorhouse.com/ and search for author=Kent.

Excerpts at http://www.bkent.net/Doc/darxrp.htm

Bill's "arguments" are very realistic;  basically, that modeling this stuff can be more complicated than it might appear (at least to some folks) on the surface.


Henry Story wrote:
Thanks for the pointer. I should point out that for every great idea
there have been great (and not so great) skeptics. These often help
advance the field, in that arguments have to be devised and clarified
to overcome their positions.
There have been people who have argued that:
  - we cannot know anything
  - the world is an illusion
  - that logic is fatally flawed
  - that the earth is flat
  - the holocaust never happened
I take it that Bill Kent's arguments must be of the better variety, but
would need more precise arguments to evaluate his position. Again
thanks for the pointer. Please do let me know if you find some
online articles that summarize his thoughts.
In the mean time, I'll just be optimistic. Logic has incredibly
solid foundations. It has received the attention of some of the
brightest minds in the past 2400 years, including Aristotle, Leibniz,
Frege, Betrand Russel, Wittgenstein, Donald Davidson, David Lewis, and
many many many more...

Henry Story

On 9 Feb 2005, at 12:00, Spork, Murray wrote:

[Cc's trimmed]

Hi Henry,

I was going to point you towards a keynote speech at Extreme Markup
Languages 2003 by Bill Kent on the "Unsolvable Identity Problem" but
can't find it anywhere - the best I can find is [1].

Doesn anyone know if it's available online? I'm sure it used to be.

Kent's book "Data and Reality" [2] is essential reading (and a little
depressing for all of us interested in the "semantics" of data) - just
as relevant now as it was in the 70s when written.

[1] http://www.extrememarkup.com/extreme/2003/keynotes.asp
http://www.amazon.co.uk/exec/obidos/ASIN/1585009709/qid=1107946684/ sr=1-

-----Original Message-----
From: semantic-web-request@w3.org
[mailto:semantic-web-request@w3.org] On Behalf Of Henry Story
Sent: Wednesday,9 February 2005 7:42 PM
To: semantic-web@w3.org
Cc: Roy T. Fielding; Atom Syntax; bloged
Subject: Madonna

Just a point of clarification about identity. (I thought the example
is fun enough that it may be of interest)

As we all know Madonna is a material girl, and she lives in a
material [snip]