ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Relating and Reconciling Ontologies

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Patrick Cassidy" <pat@xxxxxxxxx>
Date: Fri, 22 Apr 2011 17:04:32 -0400
Message-id: <000301cc0130$e0c756c0$a2560440$@com>
Do we need a "mandated" upper ontology for interoperability?
 John Sowa said;
> 
> I have no quarrel with an agreement to mandate specific
> ontologies for specific purposes.  That is the whole point of the
> microtheory organization of Cyc.
>
    If the "purpose" is interoperability, then it does indeed make sense to
have a single common foundation ontology.  But the term "mandated" is too
argumentative.  A common foundation ontology for interoperability is a tool
available for those who need to interoperate; it serves as a common language
to enable communication among the systems that use it.  It need not be
"mandated" any more than someone attending the Ontology Summit is "mandated"
to speak English - it's just that, if one wants to be understood by all the
participants, that's the language one uses.  That does not prevent anyone
from using a different language for other purposes, and translating to
English (or to a common Foundation ontology) when that is more effective in
communicating (I have heard side-conversations in many languages at meetings
where the presentations were all in English).  The beauty of having an
interlingua ontology is that it allows everyone to do exactly whatever they
consider optimal for their local purposes, and only translate to the common
language when needed to communicate with a wider audience.  It also allows
any number of specialized communication methods among smaller groups of
agents.
  The purpose of a common foundation ontology is to provide an interlingua
for translation among a very large number of other ontologies; it is only
needed for that purpose.  Conversely, without a common foundation ontology,
accurate translation among ontologies developed separately will be
impossible.  It is not difficult to show mathematically that the probability
of an accurate translation when guessing mappings from term names is
vanishingly small for any non-trivial set of assertions in different
ontologies that were mapped by some probabilistic automated algorithm.
Semi-Automated mappings can be useful provided that they are checked and
corrected by someone who fully understands both ontologies being mapped -
very labor-intensive.
   The common foundation ontology can be of modest size - I suspect it can
be kept to under ten thousand elements.  The difficulty is not in building
such a thing - the Cyc BaseKB or my COSMO ontology or even SUMO could serve
as a starting point; the problem is in gaining enough experience with actual
interoperability applications to show that the principle works, and to
supplement the starting foundation ontology when that is shown to be
necessary.
   But I am convinced that no other approach will work, and until the common
foundation ontology is given a serious try, we will continue to be having
these interoperability discussions for many years, unless the machines
evolve to human-level intelligence by some other mechanism.    (01)

Pat    (02)

Patrick Cassidy
MICRA, Inc.
908-561-3416
cell: 908-565-4053
cassidy@xxxxxxxxx    (03)

> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
> bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F. Sowa
> Sent: Friday, April 22, 2011 3:44 PM
> To: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] Relating and Reconciling Ontologies
> 
> Ali,
> 
> AH
> > Questioning the wisdom of a single, unique, mandated ontology is not
> the
> > same as arguing for a set of ontologies growing "without limit" -
> this
> > is simply a false choice.
> 
> I agree.  I have no quarrel with an agreement to mandate specific
> ontologies for specific purposes.  That is the whole point of the
> microtheory organization of Cyc.
> 
> Cyc does have an upper level ontology, but Lenat said that the
> upper level is not as important as the middle and lower levels.
> I'm not sure what he means by that in practice, but I feel it is
> an important step away from mandating a universal upper level.
> 
> And when I talk about growing without limit, I would consider the
> major growth to be in the lower level microtheories.  I would not
> expect a huge increase in the number of upper levels.
> 
> AH
> > If there are groups operating within a field with irreconcilable
> > differences in their core ontologies, do we really want to go down
> > the route of mandating that one is necessarily privileged over the
> other?
> 
> Of course not.  The example I give is physics.  Physicists agree that
> Newtonian mechanics has been superseded by relativity and quantum
> mechanics.  But they also know that for human-scale objects moving
> at reasonable speeds on earth, Newtonian mechanics generates the
> same results (within the accuracy of the measuring instruments).
> 
> Yet even for a single system, such as an automobile, there are
> subsystems for which different choices of ontology are important:
> fuel combustion inside the engine, electrons flowing inside the
> computer that controls the engine, laser-based devices for
> measuring speed and distance, etc.
> 
> Engineers are very good at making independently designed systems
> work together, even when they are based on different principles.
> Mother Nature is even more skillful:  note that the healthiest
> environments are the ones with the greatest diversity.
> 
> AH
> > Seriously, pick any field of interest. Find and count the core
> > research streams. Are they limitless?
> 
> Basic design principle:  There are only three numbers that require
> little or no justification -- zero, one, and infinity.  Any number
> greater than 1 and less than infinity must be justified.
> 
> Example:  Remember when Bill Gates said that nobody would ever
> need more than 640K bytes for the PC?  Every limit that anybody
> imposed became painfully small within just a few years.
> 
> AH
> > So, I would suggest that a more accurate choice is between a single,
> > mandated ontology ...  vs a limited handful that captures (at least,
> > immediately) irreconcilable disagreements within "unsettled" fields.
> 
> Major objections have been raised against every upper level that
> anybody has proposed from Aristotle to the present.  There is no
> prospect of ever getting universal agreement on the first option.
> 
> The second option is more promising, but the only fields that
> are truly "settled" are the ones that are obsolete.  IBM used
> the euphemism "functionally stabilized" for any product that
> they had discontinued.  My recommendation is:
> 
>   1. Adopt some version of the second option with a "practical"
>      limit that has an escape bit (i.e., like the Unicode option
>      that allows extensions from 1 byte to 2 bytes to 4 bytes...).
> 
>   2. Embed that practical limit within a framework for which the
>      only theoretical limit is infinity, but the reasonable
>      practical limit grows with the technology.
> 
> AH
> > the introduction of a novel conflicting core ontology should have
> > a high cost.
> 
> That is like saying that any scientist who comes up with a new
> theory must pay a large fee to get published.  A much better option
> is freedom of speech, but nobody is required to listen.  That's
> what we have today:  anybody can design a new operating system,
> programming language, or whatever.  But in practice, the majority
> of people prefer to stick with their old familiar systems --
> unless the new one is spectacularly better.
> 
> AH
> > ... we should be aware of the problems in the db's and sware systems
> > of the past. However, we are embarking with the explicit aim of
> > making our ontologies interoperable and reusable.
> 
> *Every* major computer system since the 1960s was designed with
> the goal of being interoperable and reusable with all other
> systems, especially legacy systems.  Those that weren't died.
> 
> The original WWW worked very well with legacy systems.  But the
> Semantic Web was designed as a total break with the past.  That is
> why the growth of the SW has been glacially slow compared to the
> growth of the WWW.  The theme of my iss.pdf talk is the need for
> a more flexible and extensible foundation for Semantic Systems:
> 
>     http://www.jfsowa.com/talks/iss.pdf
> 
> John
> 
> _________________________________________________________________
> 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
>     (04)


_________________________________________________________________
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    (05)

<Prev in Thread] Current Thread [Next in Thread>