Re: [ontolog-forum] Systems that learn each other's ontologies

["Adrian Walker" <adriandwalker@xxxxxxxxx>]

Hi John --

Interesting info about what companies have to do in order to trade with Amazon.

You wrote...

For interoperability, the goal must be to design our systems
so that they can *learn* to accommodate other ontologies in
a minimum amount of time.  A system should be able to find a
mapping from its own ontology to Amazon's, for example, in a
few minutes -- automatically -- instead of requiring a database
specialist to spend 6 weeks to write the mappings.

Well, that would be nice.  But a key question would be, how would you go about ensuring that the automated result is as accurate as the manual one?  And how would you get people to trust the survival of their company to the automated result ?

Just some late night thoughts.  HTH, 

                                     -- Adrian

 Internet Business Logic (R)
A Wiki for Executable Open Vocabulary English
Online at 
www.reengineeringllc.com    Shared use is free

Adrian Walker
Reengineering

On 7/23/07, John F. Sowa < sowa@xxxxxxxxxxx> wrote:

Arun and Paul,

Attached is a note that I sent to ontolog-forum.

It presents a solution to the Knowledge Soup problem,
which is based on some of the technology we have been
developing.  I think this is something we should write
up as a proposal to be submitted to our usual clients.

John
_______________________________________________________

Denise,

I agree.

  > There is no single way to define classes of food that works for
  > everyone and there need not be.  There are many equally valid
  > options depending on the actors and the context.
  >
  > On a practical level, not everything fits into classes.

Or more generally, anything can be classified in an open-ended
number of ways for different purposes -- just because different
people have different backgrounds, habits, and viewpoints.

An example I like to use is the Amazon.com ontology, which any
supplier must accommodate if they want to sell anything through
Amazon.com.  For any company that supplies books, electronics,
etc., that's a prerequisite if they want to stay in business.

So those companies, which usually have wildly different internal
ontologies for their databases, are forced to write mappings from
their internal schemas to the Amazon schema.  They do that, but
it typically takes them about 6 weeks to do so.

When people with different ontologies meet -- which means any two
people in the world -- they seldom understand each other very well,
and they must take some time to get acquainted, to adapt to one
another's quirks, to negotiate, to compromise, and to *learn*
how to map their ways of thinking to the other's.

For interoperability, the goal must be to design our systems
so that they can *learn* to accommodate other ontologies in
a minimum amount of time.  A system should be able to find a
mapping from its own ontology to Amazon's, for example, in a
few minutes -- automatically -- instead of requiring a database
specialist to spend 6 weeks to write the mappings.

Summary:  It's incorrect to say that people with different
backgrounds understand one another; it's more accurate to say
that they can learn to understand one another.  We can't expect
computer systems to be smarter than we are.  At best, they should
be able to learn to do something similar.

John


_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)