Re: SUO: Re: Ballot Comment
John,
At 11:39 PM 8/16/2001 -0400, John F. Sowa wrote:
>Adam,
>
>As I said many times, I don't believe that a single monolithic
>ontology is appropriate. Cyc started with one big theory, and
>later realized that they had to subdivide it into microtheories.
>SUMO will have to be subdivided eventually, and it is better to
>start work on that now rather than later.
Well, we've both made our positions clear as to whether we want
respectively a single ontology or a set of them, that's true. I don't
think it's necessarily accurate to use Cyc as an example of the
inevitability of separate theories. In the several years I worked with the
full version of Cyc, I didn't find any contradictory axioms in different
microtheories. I believe that despite the proported theoretical benefits
of microtheories for knowledge representation, that the actual use and
benefit was in constraining inference and making the search space more
tractable for proofs.
>Some comments:
>
>AP> The point of creating SUMO and putting it to a vote is exactly
> > to have our SUO "committee" critique an existing design rather than
> > start with a blank sheet of paper.
>
>We've been doing that all along, since it was on the table as a working
>document. I voted for it as a working document, but I don't believe
>that it is going in the right direction to be a standard. As I said
>at the IJCAI workshop, I would rather see it combined with IFF and
>carved up into modules than to be presented as a single lump.
That's a fine goal in theory but if you or others feel that's the right
approach then I think you have some responsibility to get specific about
exactly what terms and axioms from SUMO should be related in what way to
particular terms in IFF.
> > Some folks (including you I think, at a different time) have
> > suggested that we start with a blank sheet of paper and look at
> > "fundamental ontological distinctions".
>
>Yes, and I still think that is an important part of the puzzle.
>The fact that SUMO has not been developed along those lines is one
>reason why I'm not happy about the way it's going.
Well, then I'm confused by your point. Either we start with a blank sheet
of paper and work on those fundamental distinctions, or we critique an
existing document. Which is it that you prefer? Or do you advocate a
third option?
> > Your sentence about harsh criticism is particular apt in
> > our current case.
>
>I said that the way SUMO is going is not the way I would recommend.
>But I have suggested an alternative, which is to combine it with the
>IFF approach.
ah, but how...
> > As for implementation and testing I think as I've mentioned in a
> > previous reply to you that we're fairly far along. Not only do we have
> > SUMO "implemented" in a browser, but it's also being used in an Open Group
> > effort to develop an open source ontology for representing "Quality of
> > Service". People are already downloading that ontology.
>
>That's fine. But a browser is just an automated white paper -- it
>doesn't do anything, such as check for consistency, make deductions,
>facilitate interoperability, etc. It's nice that people are starting
>to use it, but a lot more people have been using Cyc, and I wouldn't
>vote for Cyc as a candidate for a standard.
ok, but you were citing in your last note that a reason not to support SUMO
was that it wasn't being used or tested or implemented. In fact it is.
The sad fact of the SUO effort is that for all the advice and criticism
that people have given, only a few people have been willing to be specific
enough to contribute any concrete content to a standard. Robert Kent is to
be commended for rising to the level of specificity needed to create a
standard, as should be Ian and the other folks who have contributed to SUMO.
Adam
>John
Adam Pease
Teknowledge
(650) 424-0500 x571