W3C home > Mailing lists > Public > public-sws-ig@w3.org > December 2003

Re: Cross-ontologies reasoning

From: pat hayes <phayes@ihmc.us>
Date: Thu, 18 Dec 2003 10:03:12 -0800
Message-Id: <p06001f14bc079814e1a6@[192.168.1.11]>
To: jack@networkinference.com
Cc: public-sws-ig@w3.org

I suspect this is one of those familiar cases where the, er, 
professionals, or maybe old hands would be better, know that some 
problem is almost certainly unsolvable *in general* and they know 
many nasty examples which are guaranteed to wreck any attempt to 
solve the problem *in general*, but that in practice one can do very 
useful work that pays off in real terms quite effectively, 
particularly if one keeps away from the corner cases. This situation 
tends to lead to this kind of mutually recriminationory conversation 
where the practical guys are kind of pleased with what they can get 
done, and the old hands are offended by what seems to them to be 
empty boasting.  "Works fine" for example is capable of many subtle 
nuances of meaning. To some, it sounds like 'never fails' , i.e. a 
claim to have solved the central theoretical problems. To others, no 
doubt, it merely indicates that, as we used to say in our final 
reports to funding agencies, useful results have been achieved in 
some cases.

Pat

>No, we never made the assertion that the bridge axiom creation was
>entirely automatic.  We stated that with some axioms, the use of OWL with
>inferencing algorithms can maintain the logical consistency, and answer
>cross ontology queries (which was the original question).  I think Jim &
>Bijan did a good job discussing the issues.  We do have tools that allow
>some use of algorithms that do indeed suggest bridging axioms and generate
>the appropriate OWL, but due to the different nature of these algorithms
>(probabilistic among other techniques) as opposed to OWL-DL (deductive
>logic), we advise some person-in-the loop work to ensure continuity.
>Again, once checked, cross ontology reasoning and querying works fine.  In
>addition, the precision of these algorithms is greatly increased if
>constrained within domains (i.e, not trying to compare a bookselling
>webservice with a type of cancer, but only against other e-commerce
>services)
>
>Jack
>
>
>
>>
>>  Bijan Parsia wrote:
>>>On Dec 17, 2003, at 1:24 PM, Francis McCabe wrote:
>>>
>>>>  The 'problem' I was referring to was that of automatically mapping one
>>>>  ontology (written I assume by person or persons A) to another (written
>>>>  by persons B).
>>>>
>>>>  People have asserted that there exist automatic tools for doing that.
>>>>  And I was pointing out some corner cases.
>>>
>>>For the record, I don't believe that I, personally, made such an
>>>assertion. Nor did I intend to. I didn't read anyone else in this
>>>thread as doing so.
>>
>>  I believe someone from NI made such an assertion.  Perhaps this thread
>>  started on another list and migrated.  It doesn't belong here, as Drew
>>  McDermott already pointed out.
>>
>>  -Evan
>>
>>  Evan K. Wallace
>>  Manufacturing Systems Integration Division
>>  NIST
>>  ewallace@nist.gov
>>
>>


-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Thursday, 18 December 2003 13:04:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 16 March 2008 00:10:53 GMT