Re: SUO: Re: AW: Resolution of Ballot Comments
Jim and Ian,
There is a major difference between motivation and reality.
Jim Schoening wrote:
> I understand your concern (below) that SUO-OpenCyc could at some
> point diverge from later releases of Cyc's OpenCyc. I believe Cyc's
> current motivation is for this not to ever happen, or they would have
> no reason to submit OpenCyc to the SUO WG.
It is a simple fact of life that any two collections of any kind
of information, programs, axioms, or whatever, are going to diverge
as soon as they are being developed and maintained separately from
one another.
That happens to files on the same person's laptop and desktop computers
all the time, despite the best efforts to keep them synchronized. And
it happens to every software project, even with strict controls on
updates. The only way to control divergence is to declare one version
to be "official", and every other version (which will certainly be
different) to be "deprecated". They will still diverge, but there will
only be one "official" version at any point in time.
In this case, we have three collections under different management
control: the original Cyc, which is maintained by Cycorp; the subset
OpenCyc, which is maintained by OpenCyc.org; and whatever may become
SUOOpenCyc, which would be maintained by the SUO committee under the
aegis of the IEEE.
No amount of motivation on earth is going to keep these three
collections identical. The question is whether the management of
OpenCyc.org is going to defer to (1) the wishes of the SUO committee,
(2) the wishes of the Cyc management (Doug Lenat), or (3) whatever
community of volunteers is going to contribute to OpenCyc.org.
My bet is that despite anyone's best intentions, motivations, and
willingness to defer to some designated "authority", there will be
three independent and incompatible ontologies in the Cyc family.
John Sowa