Re: [ALL, OEP] Relationship between OWL and other ontology modelling languages $swbpd

"Uschold, Michael F" <michael.f.uschold@boeing.com> wrote:

>Recently, the following email was sent to the [I think] Yahoo Semantic Web list:
>    Dear all,
>
>    Can STEP(Standard for the Exchange of Product Model Data) be used as the 
>    fundamental of industrial Ontology? If yes, how can we do? And can we 
>    use the EXPRESS schemas as  ontology discription language? If not,can we 
>    translate EXPRESS into Ontology description directly!
>
>This raises a more general and very important issue for people wanting to use 
>Semantic Web standards, but who may already be heavily using alternate modeling 
>languages. EXPRESS is just one, there are many others, (UML, FLogic, ...). This 
>group is already addressing thesauri, in a similar vein.  I propose that we put 
>this on the list of potential future Task for this group. The deliverable would 
>be a 5-15 page document outlining the general issue, and addressing each of 
>several languages with a few paragraphs or a page.  OF course, for more 
>important ones like UML, there might be a full task.
>
>Provisional name for Task: Other Ontology Modeling Languages [OOML]

+1

Personally, I find questions like "Can STEP be used as *the* fundamental 
industrial ontology?" to be quite frightening for a number of reasons.  Mike's 
suggestion addresses the language part of the above questions.  I think
that is appropriate task for SWBPD to take up.  I could help with UML and 
EXPRESS aspects of the task.

In answer to the last question, there is an open source tool which converts 
EXPRESS to OWL syntax.  It goes through an intermediate UML representation, 
loses information that all of those languages can represent, and results in 
OWL that looks like RDFS with object and datatype properties.  But it is a 
start.

Finally, we should be very careful about taking a position about the 
definitiveness of any domain or upper ontology.  STEP is a huge standard,
but it doesn't cover all of manufacturing data concerns, much less the
broader area of industrial data.  It is also not a single coherent model,
but rather a set of application area models (called Application Protocols or 
APs) which conform to a common architecture and use a common information 
language: EXPRESS.  If someone wanted to convert a particular STEP AP to OWL, 
we might get involved.  Even then, I would be careful about how the result is 
characterized.  In manufacturing, different subdomains such as Aerospace, 
Automotive, Semiconductor, Oil and Gas employ different standards to address 
similar kinds of functions and data.  This is not just an accident of history,
there are *some* good reasons for these differences.

Evan K. Wallace
Manufacturing Systems Integration Division
NIST
ewallace@nist.gov

Received on Wednesday, 14 April 2004 15:42:29 UTC