Re: Process parameter description

Of course, to determine what is "explicitly mentioned", there must be some
notion of place where the explicit mentions are expected to be.  I assume
that in practice that will be cashed out as "document identified by URI".
There's a general issue here that extends to the entire world of OWL
properties, so whatever solution(s) get(s) adopted for OWL usage in general
will likely be applicable to DAML-S properties.

Regarding the issue of what's in the profile vs. what's in the process model
- we (the DAML-S Coalition) need to arrive at a more clear statement of the
relationship between the two.  But it is clear that the process model is
meant to contain a complete statement of inputs, outputs, preconditions, and
effects - so it is not necessary for a service processing component to look
into the profile for additional IOPEs.

Regards,
David Martin

Drew McDermott wrote:

>    [Jan Ortmann]
>    Date: Sun, 25 May 2003 18:53:48 +0200
>
>    I have a question concerning parameters in the Process-Description.
>    Is there a possibility to state that a process has some parameters and
>    that these are the only ones the process has.
>
> This issue never came up, to my knowledge.  Obviously, there's a
> "closed-world assumption" at work here: any proposition not explicitly
> mentioned as an effect of an action is not altered in truth value by
> that action.  There is a long tradition of making this assumption in
> the theory and practice of automated planning.  It happens to be a
> good illustration of nonmonotonic reasoning in DAML-S, a topic raised
> in another thread.
>
> --
>                                              -- Drew McDermott

Received on Tuesday, 27 May 2003 18:17:47 UTC