Patrick Cassidy wrote:
John,
Your comments suggest that you are talking about a different kind of
interoperability than what is needed for the computer age:
On the contrary, there is overwhelming evidence that people manage to
collaborate very well without anything remotely resembling a universal
upper ontology.
. . .
people been interoperating for millennia without any common upper
ontology
Yes, *people* do, but machines don't. The sort of accurate *semantic*
interoperability that requires a common foundation ontology (or something
like it) is the ability for a *machine* to take information placed in a
public repository and properly interpret it and make important decisions
based on it. That is possible if and only if the information is specified
with the kind of precision that a common foundation ontology can provide.
If you think there is an alternative mechanism, you should provide examples
that can be *publicly* inspected and evaluated of cases where that kind of
semantic interoperability was achieved without a common ontology. Thus far
I have only seen mapping efforts which are highly error-prone and unsuitable
for automated decision, usable only to present a list of possibilities to
some human who is doing the real semantic interpretation. Examples with
details, please - not anecdotes.
My Visa card is good around the world and somehow the various banks are
able to make sense of the transaction and transfer the funds correctly
and split the fees, in spite of having lots of different accounting
systems and banking laws.
If a company in Chicago wants to rent a container in Hong Kong to ship
a load of electronics to Belgium, they can easily do it and the company
in New York that owns the container will get their rental, the company
in Hong Kong that stores the empty container will know what trucking
company is authorized do pick up the container and the yard in Belgium
will know that they can receive and inspect the container when it
arrives and will know whether the owner wants damages repaired and who
to send the repair bill to, without human intervention unless the
automated decision making requires human authorization based on
rules(Don't put more than 1/2 the value of a new container into the
repair of a used one without the leasing manager's approval - In many
cases, no ontology will convince the leasing manager that he does not
want to have the final say on the scrapping of an asset. Specially if
he is one that has to do the forensic investigation to determine who's
insurance company(trucking company, lessee, shipping company, shipyard,
etc.) will pay for it.).
The companies can all have different computer systems to control their
yards and/or logistics systems - they all understand the messages and
the rules of the trade.
If this is only possible with an "ontology" then the definition of
ontology gets pretty wide and almost every system analyst is an
ontologist.
This has gone on for years not millennia but that is only a quibble
Of course, when one person or group wants to make multiple systems
interoperate locally they can use simple domain ontologies or no ontology at
all. I emphasize again that we must clearly distinguish the problem of
general, accurate *automatic* (without a human in the loop) semantic
interoperability from all other interaction situations. The requirements
and solutions will be different for different problems.
It is not clear that anyone needs this and is willing to pay for it or
even go to the trouble of verifying it if it has no cost. A very high
percentage of the need can be met with a much simpler and restricted
solution.
If we are not clear about the problem we are trying to solve of course we
will disagree about the answer. What people have done for millennia is to
misunderstand each other and talk past (as you and I seem to be doing) each
other because of different assumptions and goals. Part of that is
inevitable because no person can know with certainty exactly what is known
by another person. Machines have the capability to be much more precise
because they *can* know exactly what knowledge other machines have, in
logical detail. I am suggesting we take advantage of that *machine*
capability by giving them the tool of a common standard of meaning with
which to share information accurately.
[JS] > This list can be repeated endlessly, but all of the real work
comes from the low-level domains.
Correction: all the detailed machine interpretation *up to now* has come
from domain-level interoperability based on direct collaboration among the
*humans* whose machines are communicating - understandably because there has
not been *up to now* any widely accepted foundation ontology. But we can
expand interoperability beyond that narrow scenario, and I suggest we look
to what *can* be done with the proper tools, and not confine ourselves to
what has been done in the past.
[JS] > Furthermore, whenever people start with an upper level, they
eventually discard it, ignore it, or relegate it to a rarely used guideline,
not as something central.
It is possible to ignore the foundation ontology only when all the
components that need to interoperate semantically are under the control of
*one* human or small collaborating group who provides the semantic
interpretations. All of your examples fit into this category. The problem
of general semantic interoperability is much broader than that scenario.
Could be true but is that enough of a justification for the expense. It
might be easier to continue to set up the standards bodies and live
with their interpretations in order to get more restricted rules that
can be verified and can be incorporated into business systems with
confidence.
Repeating again: the problem is how to achieve accurate semantic
interpretation of data when the data that is being interpreted cannot be
interpreted by some human intermediary(s) who can forge a common
understanding among separate data sources. When information is placed in a
public location to be used by others, no direct interaction between creator
and user of the data is possible and the information must be in a form that
can itself be accurately interpreted. If machines are to do the
interpretation, a standard of meaning at least as complex, expressive, and
complete as a human language must be shared. That is the role of the common
foundation ontology, to be used when there are no humans in the loop to
correct mistakes.
You can solve a lot of problems with a lot less expressiveness and the
number of mistakes becomes very small when you automate the bulk of the
process.
Pat
Patrick Cassidy
MICRA, Inc.
908-561-3416
cell: 908-565-4053
cassidy@xxxxxxxxx
_________________________________________________________________
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
|
_________________________________________________________________
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 (01)
|