Patrick Cassidy wrote: (01)
> The caveat is, that the process of waiting patiently while ontology
> applications develop and haltingly find themselves wanting to interoperate
> with others, and little by little finding commonalities that they can share,
> is at best slow and I suspect will take generations to arrive at anything
> like a usable common standard of meaning that is useful for widespread
> accurate interoperability. (02)
Well, at the rate of 1 "generation" per 5 years in the IT community, the
fact that it might indeed require 2-3 generations does not strike me as
daunting. Further, if knowledge engineering of the ontology kind
actually becomes a major part of software engineering, instead of an
academic exercise in tool building and a government experiment in
technologies for classified applications, the experience curve will be a
lot steeper. (03)
It is my impression that the existing upper ontology work has
demonstrated all of the following:
- that there are certain essentially mathematical concepts that can be
codified and shared by many practical ontologies
- that beyond that one can create several levels of upper categories
that solve no problem of themselves and are directly useful only to the
development of possibly useful mid-level ontologies that still solve no
problem of themselves.
- that the upper level categorization requires ontological commitments
that are largely irrelevant to the real problem spaces but create
serious impediments to the merger of mid-level ontologies. (04)
John Sowa will doubtless tell us that Cyc -- the mysterious and powerful
Oz -- has seen all of this and conquered it (if only they could tell
us). (And if true, it would not be the first time that a military
technology had to be rediscovered/reinvented by others in order to
become a useful technology.) (05)
What Amanda proposes is that we get some real experience using
ontologies in more than the biomedical and intelligence communities
before we leap to the conclusion that some particular Gedanken
experiment will be useful in solving arbitrary unknown problems.
And that experience is actively being acquired as we write -- the first
"generation" began several years ago. (06)
> The pace thus far suggest to me that no one
> participating in this list will live to see any widespread adoption of broad
> cross-domain interoperability by this method. That would be fine if there
> were no costs to waiting, but there are large costs. We not only lose the
> economic efficiency derivable from data interoperability, we lose the
> potential new and more powerful applications that could be developed more
> rapidly by communities that can learn from each other's results because they
> use a common standard of meaning. (07)
Well, Pat, think of it this way. The effort to create a universal
reference upper ontology with that kind of mandate will be about power,
not quality. It will be primarily governed by money and politics, not
technical excellence, and not knowledge engineering in the field. And
my 40+ years of experience warns me the result could be the Windows of
upper ontologies, and 10 years later we will have patched it into an
upper ontology that can just barely support most industrial
applications. So, if you want to risk repeating the 1990s, all ahead
full and damn the torpedos. (08)
I, for one, would prefer to see the next 5+ years spent on in vivo
testing of knowledge engineering concepts, and on the development of a
discipline. Think of it as the medical experience that will give us
some knowledge of the required properties of your panacea. (09)
-Ed (010)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 (011)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (012)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (013)
|