W3C home > Mailing lists > Public > public-swbp-wg@w3.org > January to March 2004

RE: philosophy of SWBPD (was Re: [OPEN] and/or [PORT] : a practical question)

From: Uschold, Michael F <michael.f.uschold@boeing.com>
Date: Mon, 29 Mar 2004 14:45:21 -0800
Message-ID: <823043AB1B52784D97754D186877B6CF04894E16@xch-nw-12.nw.nos.boeing.com>
To: "Bernard Vatant" <bernard.vatant@mondeca.com>, "SWBPD" <public-swbp-wg@w3.org>

Bernard says: 

"Using this expression seems to assume that
the world we try to represent has some immanent model that a 'right' representation could
somehow 'discover', while a 'wrong' representation somehow miss it. "

Well, if by 'discover' you mean increased recall in a search application, then this is EXACTLY RIGHT, Chris Welty recently wrote a paper showing that an OntoCleaned ontology gave better precision/recall than the uncleaned one. 

Most ontology applications will not have such metrics that are easily measured, and so it will be much harder to find criteria for evaluating the goodness/badness of a given ontology. Many years ago, Tom Gruber published some guidelines for building 'good' ontologies. We must not forget that goodness must always be measured with respect to the purpose to which the ontology was intended.

To shed light on what it means to mismodel something, let us identify some criteria for assessing the 'goodness' of ontology modeling choices. Note that most are far more difficult to measure than precision/recall.  Here are a few off the top of my head.
1. logical consistency - Lets hope we can all agree on this one. It can also be easy to measure, if the logic is complete.
2. various OntoClean criteria, which help to identify 
3. perspecuity: it should be easy to look at and understand a model. Some 'correct' approaches may be very convoluted, but this make them less desirable.
4. similar things should be modeled similarly, this also helps perspecuity.

We can also identify general patterns of ways that tend to commonly arise that have a low score by these criteria, and recommend these as bad practice.

Mike


 -----Original Message-----
From: 	public-swbp-wg-request@w3.org [mailto:public-swbp-wg-request@w3.org]  On Behalf Of Bernard Vatant
Sent:	Thursday, March 25, 2004 2:02 AM
To:	SWBPD
Subject:	RE: philosophy of SWBPD (was Re: [OPEN] and/or [PORT] : a  practical question)



Thanks Jim to make this point. Since I trigger the thread, should I remind that my
original question was intented to be:

- What is the trade-off in choosing A rather than B or X?
and not
- Which of A and B is right and which is wrong?

I had once a rock-climbing guide who told to begin with that he would not teach us how to
climb, that everyone could and had to find out that - IOW there is neither 'right' nor
'wrong' way to do it. What he was about was security rules and rope techniques, in order
for us to fall as rarely as possible, and if ever we do, stay as alive as possible.

I wonder what 'mismodel the world' can mean. Using this expression seems to assume that
the world we try to represent has some immanent model that a 'right' representation could
somehow 'discover', while a 'wrong' representation somehow miss it. Needless to say,
history of science in 20th century, starting with physics, has shown that this approach is
not sustainable. Even if some (many) of us have certainly various deep convictions about
what is right and what is wrong (not to say what is true and what is false), is it
possible, as a group, to be explicitly agnostic about it? IOW, is it possible to agree on
avoiding the implicit or explicit use of 'right' and 'wrong' in our future notes and
recommendations, and, to begin with, in our internal debates, and stick to engineering
issues : Does it work well, or will it screw up my system? What are the pitfalls? Which
tools can fit together? etc ...


Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant@mondeca.com

-----Message d'origine-----
De : public-swbp-wg-request@w3.org [mailto:public-swbp-wg-request@w3.org]De la part de Jim
Hendler
Envoye : mercredi 24 mars 2004 23:34
A : Christopher Welty; Jeremy Carroll
Cc : Bernard Vatant; Ian Horrocks; SWBPD; public-swbp-wg-request@w3.org
Objet : ALL: philosophy of SWBPD (was Re: [OPEN] and/or [PORT] : a practical question)


At 12:59 -0500 3/24/04, Christopher Welty wrote:
Jeremy wrote on 03/24/2004 04:24:16 AM:

>
> Yes, like Bernard, I have been thinking more about this, and about Ian's
> insistence in WebOnt that classes-and-instances was almost always raised by
> people wanting to mismodel their world. (cc Ian, wondering if I have learnt
> my lessons well!, or misrepresented him)

Well, "mismodelling their world" is not limited to classes as instances. I find it rather
dangerous to make such statements.  People use subclass incorrectly, too, but that wasn't
a reason to remove that axiom from OWL DL.  People just mismodel their worlds, I hope we
can offer some advice on both how to do some of these things and how NOT to do it.


[snip]


See, it's this kind of converse that makes me nervous -- somehow the idea that the people
who prefer separating class from instance (as Ian is quoted by Jeremy) are right and those
who prefer to use metamodeling (like Guus as quoted to WOWG. I don't have time to dig up
the mail) are somehow mismodeling.  This is nonsense -- did everyone who ever used Protege
before the OWL plug-in get it "wrong" in some sense??  Yet protege, like many other
systems, makes wide use of the extremely useful feature of treating classes as instances -
they just don't export that when you use the OWL plugin in DL mode (as I understand it).
 My big fear for this WG is that we're going to somehow "endorse" certain kinds of
representation and say other folks are somehow making errors - yet on the web, different
people with different opinions about representation will all need to use the languages, we
must be careful not to be like the "soup nazi"s in the Seinfeld show [1] who get to
dictate who gets their soup and who doesn't based on some set of rules that no one else
understands...
 Seriously, I think the BPD will do a great service if we explain the issues and the
advantages and disadvantages of various representations - but if we start to dictate one
way or the other as "correct" then we will be doing a disservice to the community and will
not be helping to deploy the semantic web.
 -JH










[1] http://members.aol.com/rynocub/soupnazi.htm
--

Professor James Hendler                   http://www.cs.umd.edu/users/hendler
Director, Semantic Web and Agent Technologies       301-405-2696
Maryland Information and Network Dynamics Lab.      301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742      240-277-3388 (Cell)
Received on Monday, 29 March 2004 17:48:41 EST

This archive was generated by hypermail pre-2.1.9 : Monday, 29 March 2004 17:48:43 EST