[MISC] Re: [MOBY-l] OWL-S, and RDF ontology for MOBY?

[Phillip Lord]

>>>>> "Mark" == Mark Wilkinson <mwilkinson at mrl.ubc.ca> writes:

  Mark> On Wed, 2004-09-01 at 10:06, Phillip Lord wrote:

  >> I think OWL-S is mostly aimed at automated composition which is
  >> not really what moby-s is aimed at. You're unlikely to be able to
  >> get semantic descriptions rich enough to describe services well
  >> enough to be able to do this anyway, in which case you are
  >> accepting a lot of complexity from OWL-S which will not
  >> necessarily buy you very much.

  Mark> We looked into this in some detail at my lab meeting last
  Mark> week.  The tentative conclusion was the we could, in
  Mark> principle, describe a MOBY service in OWL-S if we had to.  It
  Mark> may or may not be useful to do so, and we need to overcome the
  Mark> current lack of a MOBYObject->XSD translator.

  Mark> However, what we could NOT describe was the MOBY Invocation
  Mark> Message structure, since the message structure allows multiple
  Mark> invocations in a single message, each of which may be a
  Mark> different object type (so long as they are ontologically
  Mark> related).

Can you really not describe this in OWL-S? I will admit to not having
looked at OWL-S for sometime, but surely the input/output slots are
not constrained to being a specific named class? You should be able to
put in arbitrary class definitions, in which case you should be able
to use conjunctions of classes.


  Mark> So the final conclusion was that, at least as it exists today,
  Mark> MOBY-S cannot be described in OWL-S (at leats as far as we
  Mark> understand it).

  Mark> This is not a commentary on the usefulness/not of MOBY-S nor
  Mark> OWL-S, it is just an observation.


Actually, it is a commentary on both! Although, that you can't do in
MOBY-S what you can do in OWL-S and you can't do in OWL-S what you can
in MOBY-S, is probably not that surprising. Nor that much of a problem
for either. They are aimed at somewhat different markets. 

Cheers

Phil