Ergun,
This is a good suggestion, but it does not seem to address
to the root of John's question. Effort is still required, whether it is to
develop a standard ontology, or a standard for qualifying ontologies, or a set
of common, useful ontology frameworks. The funding and coordination of such
effort are still challenges.
I also
think that our previous discussions showed it would be very challenging to get
agreement on anything as a "standard", hence my suggestion that we should start
just by focusing on development of common, useful ontology kernels / frameworks.
If we can cross that bridge and actually develop frameworks that we and
others find useful, eventually one or more of them may become de facto
standards.
Without some major source of funding, it seems that a good option (and
perhaps the only feasible option) would be to follow the model
of the open-source software development community, which has been able to
successfully address the challenges of funding and coordination in many cases,
for different kinds of software and software
frameworks.
Phil
Dear all,
Maybe what can be done is to develop standards
that can actually qualify the upper ontologies developed or the algorithms
used to develop such ontologies so that they "can be called
a standard".
That way, we may be free from thinking about how we are going to build
them and just worry about the access to them.
So, the question would
be reduced to "given an upper ontology for a given domain or an algorithm for
developing such upper ontology, what is its quality / can this be a standard?"
( i.e. pertaining to its (intra/inter-)consistency and reliability).
Ergun
On 6/24/07, John F.
Sowa <sowa@bestweb.net >
wrote:
Phil,
I
agree that would be a useful thing to do:
> Rather than seek to
develop a "standard", perhaps the focus
> should be on developing
"common", useful ontology kernels /
> frameworks....
But who is
going to pay for it. If it's going to be done by
volunteers,
who is going to coordinate the volunteers, and
how do you ensure
consistency, reliability, etc.?
As for systems that people have paid
money to build, Cyc is
the premier example. But after 23 years
of development, it
still does not generate enough money to keep itself in
business
without a major influx of research grants.
Over the
years of SUO efforts, we went round and round on these
funding issues
without getting answers that had money
attached.
John