Saturday, November 19, 2005
Center of Excellence Proposal
à
Barry
Smith:
There is
visibility, interaction, effect in the world of drosophila,
for
example. There they seem not to lead to service descriptions or policy.
Paul S
Prueitt:
I am not
sure how you feel about the OASIS draft (attached) Reference Model for Service
Oriented Architectures.
Barry
Smith:
A
top-level ontology is designed to create the possibility for
interoperability
between many lower-level domain ontologies. It is
not designed
to substitute for the latter.
Paul S
Prueitt:
A
stratified ontology (as I have suggested) should have templates or specific
types that assist the real time aggregation of atomic elements (parts of things
with variable meaning depending on circumstance/context) into models of a
specific situation/event.
Barry
Smith:
I doubt
that policy, or service description, are atomic elements. Do
you have
a definition of 'atomic element'?
Paul S
Prueitt:
The
concept of a semantic primitive is not precise, but in some cases I would
suggest that semantic primitives can be aggregated into a model of an
event. The model would then be in
correspondence to the reality of the event, and the concepts - the set of
concepts involved might be called an ontology, or better "an ontological
model" of the event.
In my
reading of C S Peirce, he felt that a concept was like a chemical compound - it
is formed from an aggregation of "atoms". The key, in the opinion of those who think about this like I do,
in understanding this event chemistry notion is to reflect on the real physical
different between quantum reality and chemical reality.
Barry
Smith:
At some
point physics, biology, chemistry, etc. needed to
move
beyond the point where games with loose rules that could be
played
with little effort to find the truth. It is the thesis behind
the
foundation of NCOR that ontology has reached that point.
Paul S
Prueitt:
NCOR
stands for the new National Center for Ontology Research, which I take it you
are one of the founders.
I do not
feel that you are understanding my notion of "loosely held".
I know
that there are many things that can be engineered, many "web
services". So this is not where
the issue of stratification comes up.
If there is no emergence there is no need to move away from a purely
Newtonian model.
The
concept of "loosely held" might be better stated as
"underconstrained"... as in an emergent process involved in
neurochemistry might be underconstrained if it were not for the actually presence
of specific reactants in the environment of the process.
I have written about this for a decade.... but have to admit that not many individual follow what I have been saying. Some do.