Re: Draft OWL-S Process technical ['securiQ.Watchdog': überprüft ] ['Watchdog': checked]

B) is close to the interpretation I intended, except that I would say that 
'InCondition' was true AFTER (at the end of) service invocation.
Effectively, the client can deduce that InCondition is true based on the 
message received.


At 05:07 PM 7/6/2004, you wrote:

>hi Drew and all,
>
>i have a question regarding the "InCondition"/"WithOutput"/"hasEffect"
>construct:
>
>How should an agent interpret such a construct...I think there could be
>(at least) two possible interpretations:
>
>a) "if 'InCondition' is true, then expect effect 'hasEffect' to occur and
>expect to receive an output 'WithOutput'" (somehow an 'ex ante'
>interpretation)
>
>b) "if you receive an output as defined by 'WithOutput' then you can
>assume that the effect 'hasEffect' has occured, which means that the
>Condition 'InCondition' happened to be true on service invocation"
>(somehow an 'ex post' interpretation)
>
>In the context of the credit card example, the difference may be
>negligible, because both the service and the agent may have the means to
>check the condition, i.e. the limit of the credit card. However, the
>question may arise for services that are fully nondeterministic and make
>opaque decisions, e.g. a lottery service, where the agent can not figure
>out the condition in advance, which would mean that interpretation (a)
>would be not adequate.
>
>(i apologize if this was already discussed in ealier mails)
>
>thanks and kind regards,
>Joachim
>
> > With some trepidation, I've attached the draft of the new process
> > technical overview for Release 1.1 of OWL-S.  Comments welcome.

* Dr. Mark H. Burstein <burstein@bbn.com>
* Director, Human Centered Systems Group
* BBN Technologies
* 10 Moulton Street
* Cambridge, MA 02138
*
* Tel: 617-873-3861
*  Fax: 617-873-4328
* http://daml.bbn.com/burstein