W3C home > Mailing lists > Public > www-webont-wg@w3.org > May 2002

Why paradoxes would render OWL useless

From: Jonathan Borden <jonathan@openhealth.org>
Date: Wed, 22 May 2002 17:12:16 -0400
Message-ID: <007901c201d5$5a65fb90$0a2e249b@nemc.org>
To: "Jonathan Borden" <jonathan@openhealth.org>, "Dan Connolly" <connolly@w3.org>
Cc: "WebOnt WG" <www-webont-wg@w3.org>
> The real world fact is that we have to make some decisions in relatively
> short order else we might be stuck with something that, if paradoxical, is
> useless.
>

The reason I make such a strong statement is from personal experience. In
1983 I was involved in a project that collated experimental results of
chemical reaction pathways in the brain -- involved in Parkinson's disease
as well as calcium ion channel driven neurotransmitter release -- into a
graph structure which had attached to it various axioms and equations. The
"system" ran a simulation in a sort of Petri net fashion. The program took
months to run and produced results. Some of the results were surprising,
remarkable even. Going back over the network and the equations I couldn't
figure out how these remarkable results would arise from the axioms and the
initial state of the system.

I just couldn't tell whether my results were real or artifacts of the system
and hence I didn't publish (despite the urging of my professors). A couple
of years later someone ran a sophisticated experiment and demonstrated in an
observable fashion, the results that my software predicted. I've had regrets
to this very day.

If only I could have trusted my software ... ***

Yet I am convinced that the "Semantic Web" will produce results that we will
need to trust, and hence we really need to get these issues right.

I have no doubt others in this group have similar stories but just to point
out that I don't think I am throwing around hyperbole.

Jonathan

*** I called the language "SENSE" for "SEmantic Network Simulation of
Events" :-))
Received on Wednesday, 22 May 2002 17:16:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:50 GMT