[wg11-owl] [Fwd: RE: Engineering uses of OWL]

[Ed Barkmeyer]

Folks,

I sent out a few probes, and this is the only positive response thus 
far...

The PDTnet response was essentially: "we model what we need to capture, 
and we leave the 'reasoning' to data mining tools."  That, and the 
mention of Fair-Isaac (Blaze) below, suggest that datalogic/SCLP, rather 
than FOL or DL, may be the preferred reasoning technology for 
engineering (as Evan hinted).

"Situated" logics, like datalogic, reason from current knowledge 
(facts), rather than "ontological truth".  The main difference is in the 
interpretation of NOT X.  In datalogic, NOT X means "X is not known to 
be a fact by the agent at this time (even after any deductions the agent 
was told or knew to perform)".  In mathematical logic, NOT X means "X is 
known or deduced to be always false."  (Benjamin Grosof described the 
difference as "you must have pure mathematical logic, or ELSE.")

-Ed

-------- Original Message --------
Subject: RE: Engineering uses of OWL
Date: Mon, 6 Jun 2005 10:49:06 -0500
From: Burkhart Roger M <BurkhartRogerM at JohnDeere.com>
To: edbark at nist.gov
CC: Smith Dave C <SmithDavidC2 at JohnDeere.com>

Ed--

The Knowledge Based Engineering (KBE) initiative in MANTIS has on its
roadmap the representation of engineering design problems in knowledge
forms that could be processed by rules engines.  Their current RFP,
about to be issued, focuses on services for KBE interoperability
rather than specific engineering design models.  At least one of the
business rules vendors (FairIsaac) has indicated that customers are
applying commercial rules engines to engineering design problems.

In general, automating the generation of designs, or the processing of
design knowledge for multiple kinds of analysis, are candidates for
semantic knowledge representation and inference or other processing.

Collaborative Product Design Associates (CPDA, the successor firm of
D.H. Brown) is looking at a pilot program for a "CAE data model" which
deals with representation of engineering designs for CAE processing.
They're working with various players in the CAE industry, and held a
recent workshop in Atlanta that included discussion of shared semantic
models to integrate various players.  See the description of their
workshop "Developing a Design/Simulation Framework" on their web site
at cpd-associates.com for more information.  I'm continuing to talk
with them, but most of the companies they have involved seem to be
going more down their own paths rather than any use of SC4.

I don't see any reason not to eventually move all engineering-related
information to forms that can be processed for semantic reasoning.
The work in OMG in mapping OWL to UML metamodels, and your work and
others on UML and MOF versions of EXPRESS, should eventually build
enough bridges so that all engineering information can be turned into
OWL or whatever form is needed to enable reasoning.  I did some
experiments last year with David Price of Eurostep which turned XMI
exports of a SysML structure modeling prototype into OWL for viewing
under the Protégé tool, but we didn't get to a point of doing any
further processing against the OWL representation.

At Deere, I can't say that we've yet "crystallized some problem
categories that you expect reasoning technologies to solve," but
design and analysis automation is a general category in which I
expect they'll apply.

--Roger


-----Original Message-----
From: Ed Barkmeyer [mailto:edbark at nist.gov]
Sent: Wednesday, June 01, 2005 12:31 PM
To: Roger Burkhart; Smith Dave C
Subject: Engineering uses of OWL

Gentlemen,

SC4 has commissioned a small group of "interested parties" to determine 
whether/how the semantic web language OWL might reasonably be used in 
the SC4 world.

What I am asking is whether either of you knows of an "engineering 
information analysis" problem that might be solved by a reasoning engine?

Deere seems to be pretty forward-looking, and I'm hoping you folks have 
crystallized some problem categories that you expect reasoning 
technologies to solve, and may well have experimented with.

If there is anything (or person) you can point me to, it would help.

David Loffredo at STEPTools is responsible for overseeing the SC4 
effort, and his intent is to leave to the OWL apostles the finding of 
disciples with real problems.  I'm just contacting a few known 
associates who understand the technology and still have their feet on 
the ground.

Thanks,
-Ed

P.S. It occurs to me that "feet on the ground" has a somewhat different 
interpretation in David's high-tech agriculture pursuits.  ;-)

-- 
Edward J. Barkmeyer                        Email: edbark at nist.gov
National Institute of Standards & Technology Manufacturing Systems 
Integration Division
100 Bureau Drive, Stop 8264                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8264                FAX: +1 301-975-4694