Cory, (01)
I agree with you. That is why we have to focus on methods
for finding "hubs" or "cores" and relating them to one another
and to the more detailed problem-oriented special cases. (02)
That's the whole point of the lattice of theories: it's
an infinite collection that lays out all possible hubs,
cores, subsets, supersets, etc., together with all possible
generalizations, specializations, and combinations of them. (03)
Since many people are overwhelmed by infinity, I take a step
back and say that we will only put those theories that people
have actually implemented and used into the metadata registry.
But we will also include the links that show how those theories
are related to one another. And those links are the same four
operators that are used to relate the theories in the infinite
lattice: contraction, expansion, revision, and analogy. (04)
> So, the pursuit of a "local hub", which is as "universal"
> as we can make it, may have substantial benefit even if it
> is not the one, true, cosmic and universal answer. We don't
> need agreement outside of this community; we need success
> at providing a compelling capability in a supportive community
> ecosystem. (05)
Yes indeed. Once we accept that point, the major issues to be
discussed are: how big is the core, how many hubs will there be,
how are all these things related to one another, and most of all,
how do they relate to legacy systems, which will coexist with
the brave new world for many years, if not decades. (06)
The biggest battles arise when people want to put their own pet
theories into the core. Murray Altheim, who has had a great deal
of practical experience in these matters, listed the following
five requirements for the core: (07)
> 1. a means of establishing [subject] identity
> 2. a facet, or property relation
> 3. the class-instance relation
> 4. the superclass-subclass relation (which I later redefined
> to be based on a mereological or collection basis, ala Cyc)
> 5. a means of expressing context (08)
I would choose different words for those five things, and I'd
be willing to add a few more. But I'd keep the core very small,
much smaller than most people are asking for. I'd also recommend
more than one hub and insist on systematic methods for relating
hubs to one another and to the problem-oriented theories. (09)
And there's still one *very big* issue that hasn't been adequately
addressed: how do the concept and relation types in the ontology
relate to the words and phrases in natural languages. (010)
John (011)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (012)
|