W3C home > Mailing lists > Public > public-awwsw@w3.org > June 2009

Last AWWSW...

From: Jonathan Rees <jar@creativecommons.org>
Date: Thu, 4 Jun 2009 17:23:49 -0400
Message-ID: <760bcb2a0906041423j7f5c42f2x6898e395b6e827e6@mail.gmail.com>
To: Harry Halpin <hhalpin@ibiblio.org>
Cc: AWWSW TF <public-awwsw@w3.org>
I felt really bad that things degenerated to the point that you would
say some of the things you did... quoting you:

"the problem is that are unresolvable differences re ground rules. and
everytime I try to interact with this group I get into far more
disagreement re ground rules than practical engineering. because if
you believe everything is verifiable, that's fine by me. I think
that's wrong. My goal is not to convince alanr that logical
verificationism is wrong or to try to convince dbooth that
Tarski-style interpretations have almost nothing to do with the world
besides constraining inferences."

First of all, the only ground rules we've articulated are listed at
the top of this page:
Do you disagree with these? If so that would be an interesting thing
to talk about and it would be sad if didn't want to contribute to a
critique. I think many of the questions on the call that got us into
trouble were attempts to constructively test your work by applying
these rules. I don't think describing them as "verificationist" is
accurate, although I can see why you applied that caricature.

Of course it might be very helpful if you *could* convince Alan to
loosen up or David to use a more mainstream approach, even if that is
not your goal.

My take on ontology quality is that any ontology is the rule-book for
a game that people might engage in. Moves in the game consist of the
publication of statements that use its terms in response to the
conditions they find themselves in (that includes previous statements
they've read, observations they've made, conversations they've had
with people, and so on). Sure, terms will acquire meaning through
play, but isn't the game more fun if you can predict which moves are
going to be OK and which will lose you points? The problem with
vaguely specified terms - games of chance, basically - is that I'm
afraid that if I use one, some authority will jump out from behind a
tree and say "you can't do that" - or equivalently that it will fail
to be of any use to someone reading what I write, or will lead to
disappointment in data integration because someone else has used it
inconsistently. These have all happened to me and I really don't like
it. With soccer you're at the referee's mercy, but only to a limited
and predictable extent. The games we invent ought to be as good as
soccer. Bad rules, bad game.

The method Alan uses of definitions and what he calls "realism" is
just one way to create rules. I find it a bit limiting and pedantic,
but he has employed it with success in several ontology efforts. If
there are other effective ways to prescribe applicability of a term, I
am very interested to learn them. (E.g. Pat has said there is
something called "knowledge representation" that uses an axiomatic
approach instead of definitions. Another approach might be "emergent
meaning through use" where you toss a term out and wait to see how
people use it, and then that becomes how you're supposed to use it.
Trademark licensing gives another approach - the contract spells it
out, and a judge or jury decides if there's a dispute. "URI ownership"
might be a good method if there is clarity over the implicit contract
connected with the URI. Examples where any of these approaches has
worked would be very interesting...)

So regarding a review of IRW, I think the question is, what kind of
review would you like? What are your goals? How would you like the
ontology judged? What do you think makes a good ontology? Instead of
applying standards you don't like, perhaps we should just try to meet
whatever needs you have.

Regarding AWWSW, the question is, how would you answer the various
riddles we're trying to cope with, as recorded in the wiki pages and

Received on Thursday, 4 June 2009 21:24:30 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:07 UTC