Re: SUO: Different questions then.
On Wed, May 22, 2002 at 12:21:27PM -0500, Richard Martin wrote:
>
> Suppose I have three operational systems. System A has some kind of
> ontological engine that adheres to the methods allowed by the
> Skolem...Godel variety of logic, system B has some kind of ontological
> engine that adheres to the methods allowed by the Zermelo-Fraenkel
> variety of logic, ...
I'll assume by "variety of logic" you mean "variety of set theory". And
by "Skolem...Godel" I assum you mean "von Neumann, Bernays, Godel".
> and system C needs to mediate the interoperation of system A with
> system B. What kind of ontological engine does system C need to
> support this mediated interoperability? Does the ontological engine
> of C even make a difference for the purpose of achieving the
> interoperability of A and B via C? Can A and B interoperate completely
> without C?
I'm not quite sure of what to make of the first two questions.
Regarding the third, since VNBG is a conservative extension of ZF,
(the language of VNBG differs from ZF by the addition class quantifiers)
every theorem of ZF is a theorem of VNBG. So anything you say in the
language of ZF is transparent between A and B. Statements of VNBG
involving class quantifiers don't in general translate into ZF, so there
is no general mechanism for integration from A to B.
> Replace the "... variety of logic" in A and B with 3D and 4D data
> representations and ask the same questions about C. Are the answers
> similar?
Probably not, as 3D and 4D are "competing" ontologies. VNBG, by
contrast, simply enriches the ZF ontology. Everything that exists
according to ZF exists according to VNBG. How to integrate competing
ontologies is a matter of some controversy, to say the least.
> Put another way, consider a fashion studio that needs new equipment
> for production and the term 'wear' appears in the purchase information
> exchange. How does the ontology of each distinguish the meaning of the
> term when encountered in the information content of the other? If I'm
> the equipment broker, how does my ontology keep the meaning straight?
That strikes me as a rather more straightforward problem than the 3D/4D
problem, and seems quite easily solvable by keeping namespaces straight.
"wear" is simply ambiguous across the two ontologies and has to be
disambiguated by some mechanism to make integration possible, e.g., a
prefixing mechanism: "PRODUCTION::wear" vs "PURCHASING::wear" (or
whatever the two relevant ontologies are here). Whether the same sort
of mechanism can solve the 3D/4D integration problem is, again, a matter
of some dispute.
-chris