Dear Matthew, Ali, Chris, et al., (01)
You have made thoughtful, well-reasoned responses to the various
issues. But I agree with the spirit of Ali's closing remark: (02)
AH> I was hoping to see how the 3D-4D debate has practical
> implications for ontology interoperability. I have difficulty
> pinning down these problems. (03)
If you try to do a global alignment of two very large and detailed
ontologies, you would certainly encounter many difficulties. (04)
But the point I have made repeatedly is that interoperability
between two or more computer systems is a much, much simpler
problem than doing a global alignment of all their ontologies. (05)
1. Two independent systems X and Y interact by passing messages. (06)
2. The only ontology relevant to their interoperability is the
subset that includes only those terms (words or symbols) that
actually occur in the messages that pass between X and Y. (07)
3. Any aspects of their global ontologies that do not occur in
the messages are irrelevant to interoperability between X and Y. (08)
For example, System X might have a 3D ontology and System Y might
have a 4D ontology. The ontology of X might treat dates and times
as designations of points in a one-dimensional flow, but Y might
treat the same information as one component a 4D coordinate. (09)
The axioms for System X or System Y involve additional information
that is incommensurable with the axioms of the other system.
However, there are widely used standards for specifying times and
dates that are independent of a 3D or 4D ontology. (010)
Those standards use a small subset of the information specified in
the internal ontologies of X or Y. But any internal information
that does not occur in the external messages is irrelevant to
the interoperability of X and Y. Therefore, the small ontology
of the messages should be the basis for interoperability, not
the much larger internal ontologies of either one. (011)
I'll acknowledge that if X and Y involve space ships traveling
through the galaxy at very high speeds, relativistic effects
might create discrepancies that are not accommodated by the
current standards for dates and time. In that case, the
messages would have to include more information, but they
would still not require a global alignment of all ontologies. (012)
Summary: Interoperability among systems should be defined
in terms of the ontologies of the messages that pass among
those systems, *not* in terms of their internal ontologies. (013)
Systems designers have used that approach for years, and it
works for current systems and legacy systems. For future
systems, it is likely to be more flexible and extensible
than any universal global ontology that anyone could hope
to develop today. (014)
John Sowa (015)
_________________________________________________________________
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 (016)
|