SUO: Re: Industry takeover
John,
At 10:44 PM 4/29/2003 -0400, John F. Sowa wrote:
>Adam,
>
>Nobody was criticizing the charter.
>
> > All we have to do is propose something that meets
> > the charter, not justify the charter itself at
> > this point.
>
>I was simply observing that you were requiring more
>proof and justification for IFF, which has already
>been approved as a candidate project, than you were
>requiring for SUMO.
I disagree with that, but let's move on.
>I believe that all three projects,
>IFF, SUMO, and OpenCyc are worthy candidates.
Glad to hear that. Do you mean that literally, or should we add the
qualifier "...if combined into a single proposal." ?
> Since
>IFF has already been approved, no further justification
>should be necessary.
Quite true, but evidence as to the benefits of some sort of merge is
necessary, if you expect people to work to that goal.
>Your next paragraph makes a good point:
>
> > You'd like SUMO to collaborate in some way with IFF.
> > That's fine. But even putting aside for the moment
> > whether I think that's a good idea, you've got to
> > provide some concrete example of what you want us to do.
> > I really have no idea concretely how to begin to do what
> > you ask. Could you do the first day of work on what
> > you'd like to see us do so we'd have a model to follow?
>
>That is a fair question.
>
>Tomorrow, I'll send a suggested outline of how developers
>from the three efforts could collaborate, but I would also
>like to ask Robert and the other IFF developers to comment
>on this point.
I'm hopeful about this, but let's be clear. This wouldn't be an outline,
or a discussion, but a little bit of actual work toward the sort of merger
that you are proposing. If morphisms are a good thing, define a morphism
from some theory in SUMO to a theory in Cyc for example, on the basis of
IFF and provide an example that supports how that morphism can be used on a
practical task. Could it be used on converting some ground statements in
one ontology to the other for example? If so, select a few ground
statements, perform the morphism and explain why a simple thesaurus won't
suffice. Or explain why a "lifting axiom" defined in first order logic
isn't sufficient, or as efficient, for a more complex mapping. This can be
a job that takes less than a day, if the benefits are as clear as you
promote them to be. Then we'll have something concrete to debate, or maybe
nothing to debate at all since the benefits will be clear if the work is
successful.
Adam
>John
>