Lists Home |
Date Index |
- From: "Bullard, Claude L (Len)" <firstname.lastname@example.org>
- To: Uche Ogbuji <email@example.com>,Martin Bryan <firstname.lastname@example.org>
- Date: Tue, 19 Dec 2000 13:24:53 -0600
Asking too much is part of getting a good deal.
Write an ontology for buying a used car. :-)
It means what you predict. Ontologies are
self-fulfilling internally (consistent) and predictable
when used to service multiple parties (coherent).
Dictionary + Tests. Services are tested. The
axioms can be behavioral. We probably should
do a little more work understanding the original
author's meaning for the "knowledge level".
It seems however, that the arbiter of commitment
is test results per an interpreted description.
That corresponds reasonably well to how contracts
that depend on standard references are negotiated.
Perhaps defining tests appears to be outside
the mission of RDF or Topic maps, but I think it
is just another ontology and set of resources.
We may need an ontology to proof an ontology
and we are back to compressability, perhaps, a
topic map of RDF assertions.
So, a question to ask is what are the tests that
could be applied to ensure the user of a vocabulary is using it
correctly without requiring them to answer any
arbitrary query? Note in the original source
the comment about not requiring completeness
or requiring the global users to share a theory.
It seems reasonable to test the existence of
assertions. We can look at multiple ontologies
(contexts of assertions) and discover that multiple
sources have the same assertion so establish
evidence by multiple sources. We can never
formally prove the assertion unless we both
agree to a test of fact and commit to behave
accordingly in a testable way.
George W. Bush is the president-elect of the
US. That is easy to test. George Bush is
the legitimate president of the US. That is
harder to test. George Bush will be a good
president. That is impossible to test until
we agree on the conditions to be met in the
No, it doesn't end. Like a recursion, we tell
it how deep is deep enough. The cabala ascribes
the aleph in the golem's mouth for a reason.
Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h
From: Uche Ogbuji [mailto:email@example.com]
I do think you and Len ask too much, but oh well.
This is "a clear definition of terms"? Whatever, let's not get into a
battle over the meaning of "meaning". I'm happy to seize on the
mention of "human-readable text" as you do.
RDF recommends the use of human-readable "label" tags
in schematics. So at least the meta-meta data are described as you say,
which is, I think the key part.