RE: XG W3pm Scope

Dear Henson,

Just to say that I agree the below but want to ask you/all to keep firm statements in this stage limited:

example1:
when you say:
A constraint for a PM ontology language is 
I would say: A constraint for a PM ontology language could be....(we are not sure yet)

example2:
when you say:
we should (1) document use cases & requirements for, say, n-ary datatype predicates;
I would say: I am not sure at all that we SHOULD do this (yet)...

well you get my point....let's stay in the beginning at the right abstraction level with complete open minds without have to much, maybe inappropriate assumptions.

(no critisism...I did the same when I described QCRs when talking decomposition.....)

Cheers, Michel

ps
keeping the above in mind I think there is on the other side room for playing around at solution side too: I attached some sketch of David's ideas (part of) for which I still have quite some questions (?do we need subclass of owl:Clas really, how to extend for derived quantities, how to extend for prefixes, how to position calculus stuff as indicated etc.etc.). The picture btw quite conforms to the idea of a "product ontology language" as described by you.

ps2
sorry for the attachement, as soon as w3c has provided the wiki we can put all stuff there.....


 
I agree with the comment "we need to know the purpose for creating these
models in this XG. If a primary reason was to support reasoning with
Description Logic tools that affects what and how you define things in
your models."  

My provisional answer to the ultimate objective of the XG is to provide
a Product modeling ontology language with perhaps a number of reusable
ontologies (the modules) for this domain. The models are perhaps low
hanging fruit, provided we have a big picture and know how the modules
fit into this picture. So we need to "know how people from our community
want to use the language."  I believe that we can get to a big picture
in a finite amount of time by reviewing use cases, requirements, as well
as listing constraints and ground rules.  

A constraint for a PM ontology language is as Kendal says "that OWL be a
good fit for product modeling, but that we try to stay within OWL DL,
chiefly OWL2 DL (or whatever OWLWG ends up calling it)."  ... Certain
design decisions will make DL compatibility quite difficult, and some of
those choices were already suggested today!" I agree with the comment
"we should look at which useful DL subsets there are that can be
exploited by various tools" Product modeling use cases almost certainly
extend current DL.  We of course need to write these use cases down and
examine them carefully. DL and related tools are evolving (Motik's
structured objects and integrity checking, for example), and we need to
keep this in mind.  

To paraphrase Kendall we should (1) document use cases & requirements
for, say, n-ary datatype predicates; and (2) make sure that OWLWG knows
what they are so that we and the OWL group maintain consensus and
awareness, inasmuch as possible in order to make our measurement & unit
"stuff" fit what OWL2 will provide.

The use cases for quantity are not as simple as they may first appear.
For example, in working with Ian Horrocks on applications of OWL 1.1 to
Systems Engineering we have had occasion to model product requirements
such as weight requirements, e.g., that the weight of an air vehicle
product is < 33000 pounds.  We have represented weight as a "quality"
class which is dependent on the product class.  Weight-in-pounds could
be defined as subclass of this, perhaps through an existential
restriction on the property value "pounds" in a quantity class.  In any
case only this kind of stuff needs to be worked out carefully to avoid
perpetrating unusable solutions.   

- Henson


This e-mail and its contents are subject to the DISCLAIMER at http://www.tno.nl/disclaimer/email.html