RE: Cross-ontologies reasoning

Charles White wrote:

>I'm not sure that you would use ontology mapping for factory floor control 
>(at least not in the near future). Would you trust someone else to send an 
>ontology in to your factory to control anything?

I think that the factory floor example was being used to describe one
end of a spectrum of mapping tolerance.  Manufacturing control is an 
application where a mapping error could endanger people, equipment, or 
product (listed in decreasing order of importance, but even the last is 
generally unacceptable to a manufacturer).  Yet there are still multiple 
ontologies that must be mapped, say from a planning system to an execution
system to an equipment controller.  These ontologies will also change, albeit 
under controlled circumstances.  One way they might change is via a remote 
vendor diagnostic session or upgrade.  So, yes, a trusted organization may 
be allowed to "send" a new ontology to a control system on a factory floor.

The mappings to this revised system would certainly have to be verified 
before they would be relied upon to control manufacturing processes.  In 
such an environment, minimizing the downtime for these changes while 
maximizing the confidence in the mapping is critical.  If SW based 
languages and tools can help with this, then they are of value even
though a manufacturing environment isn't as chaotic as the web at large.

-Evan

Evan K. Wallace
Manufacturing Systems Integration Division
NIST

Received on Monday, 5 January 2004 16:46:28 UTC