Monday, January 16, 2006
[350] ß parallel discussion in “national debate” bead thread
This is part of a discussion
that will be moved to a Wiki page soon.
Ontological modeling with
category theory and convolutions
Post OWL strategy
[347] ß Judith Rosen’s recent communication
(In reference to an email from Dr Richard Ballard, founder of Knowledge Foundations Inc).
Dick,
Thank you for you kind note, you continue to be part of my
family of friends.
I do not know if the change to talking about my work will lead anywhere, but it might. I am grateful for virtualTaos investments and for the new contract to ontologystream for consulting services.
I have had long periods in my life where I was only
concerned with the deep theory that I now have, and have in written form. My being in Taos feels like a welcome and
safe spot. I am building new computer
programs now. The architecture is
consistent with your Mark 3 architecture.
I, like you, have developed an extensive description of
work that is almost characterized by "me", even though I feel I have
taken the time to ground this work in the work of others.
The note to the BioPAX group
opens up a path to them/us thinking about "what is next". I will be interested in the response, but
can predict it.... they are properly busy doing something well defined; and the
note from me is about the step after ... they complete this important first
step. But they are natural scientists,
not computer scientists, and as a consequence they and I talk the same basic
language.
The "ontology" community is so absorbed in the
technical difficulties of OWL and RDF standards that few are able to step back
and realize the validity of the specific objections to specific foundation
issues (unique IDs, and URIs) as well as the limitations of logic and mathematical
formalism (when applied to natural complexity). The ignorance of these issues, in the computer science community,
are leading to a barely functional semantic web. But when compared with world wide data interoperability now,
this barely functional semantic web
will become a true economic blessing.
The completed Semantic Web will "work" but only
as a oddity that does appear to (occasionally) do all the wrong things at the
right time. So there will be profound
train wrecks in the use of Semantic Web realities. But the emerging Semantic Web will, as Mill's points out,
automate e-business in a way the is worth a few trillion, over the next ten
years, in gained productivity. His work
with Knowledge Foundations Inc is most important.
But
the simplified general theory /
architecture reveals a strategy that I feel can be implemented next,
given that BioPAX and others and many business modeling ontologies are actually
being developed to be read by the Stanford Protege software for editing and
using OWL.
What Protege does NOT allow is an export to a pure concept
ontology having none of the features that make Protege difficult to code. However, I can see how to use a web portal
to harvest well-formed Protege (an entanglement of Java and RDF/OWL) and stand
the "content" up as a language independent Topic Map. This could be done as a service to the
larger community that is trying to understand what ontology is and how it can
be used to standardize data schema. It
has to be far simpler than what the Stanford team has so far made
"it".
Here enters the OASIS SEE and SOA standards. SEE stands for Semantic Execution
Environment, SOA stands for Service Oriented Architecture. SEE tries to address the need for
flexibility in creating a situational finite state machine that arbitrates the
data exchanges between two computer programs that might not have the same set
of concept representations - and intended interpretations of those concept
representations. Example: an order for
parts is a concept that can be represented using different data schema, and
work flow (order completion). Two
companies will instrument order fulfillment using different standards (social
agreements) because there is no correct one way to do the representation. So the mediation process is needed, to orchestrate
an iterative alignment of the intended meaning of the "web
service".
My current contract may allow the development of a hybrid
standard called SIE (Semantic Interpretation Environment). Here is where one hits the religious nature
of the W3C advocates. They do not
acknowledge the need for a human in the loop and the notion that computer
programs should "know" when they to not "know" correctly
(with the definition of "correct" being "as humans intended the interpretation
to be".)
So we are setting up a migration path that uses the
expected hard instrumentation of human communication about data systems into
OWL, and into Protege in particular.
This expectation is setting up both the proof that logic is limited and
the data resource that can be mined and converted into simple n-aries.
We, you and I and a few others, will then be able to show
the alternative as being both far simpler that anyone will allow now, and far
more useful. The automation will be
coupled with human in the loop where the human will be non-computer scientists.
Mill's vision is consistent with the SIE pathway, but he
does not as yet see the details that you and I see about how the Knowledge Age
will suddenly appear.
In many ways we are already there, but the emergence of
the OWL ontologies will be the last design mistake (it is my expectation)
before true knowledge based tools will be ubiquitous in education, government
and commerce.
Dr Paul Prueitt PhD (mathematics and quantum cognitive neuroscience)
The Taos Research Institute
Taos New Mexico