W3C home > Mailing lists > Public > www-ws@w3.org > May 2003

Re: DAML-S: nondeterministic services

From: Mithun Sheshagiri <mits1@cs.umbc.edu>
Date: Sun, 04 May 2003 15:30:40 -0400
Message-ID: <3EB56A60.80701@cs.umbc.edu>
To: Mark Burstein <burstein@bbn.com>
CC: www-ws@w3.org

Thanks for your reply Mark.

Mark Burstein wrote:

>Compositions of such a service would have to account for both possible
>outcomes being returned. Obviously, from a planning perspective, one is the
>desired outcome that would cause the service to be selected in the first
>place. 
>  
>
            This is precisely how I use the service as a operator in my 
planner. What I would also want to know is the best way to model such a 
service in DAML-S and how the description affects the interaction 
mechanism between the user of the service and the service provider.

peace,
Mithun

>
>Mithun Sheshagiri writes:
> > 
> > Hello all,
> >                 Suppose we have a service ApproveCredit that takes in 
> > CCName, CCNumber and CCExpiryDate as the input and produces the ValidCC 
> > and InValidCC as the conditional effects based on the condition VALID. 
> > If this is the right way for modeling this kind of service then how 
> > would you advertise this service.
> > 
> > The service cannot indicate the "value" of VALID if inputs are not 
> > known. This service is nondeterministic w/o the value of VALID. How will 
> > you handle such services for composition. One possibility is to have a 
> > 2-phase method. The first phase involves providing the inputs to the 
> > service provider who in return will provide the value for VALID. Once 
> > VALID is known, composition is possible.
> > 
> > peace,
> > Mithun
> > UMBC
> > http://www.cs.umbc.edu/~mits1
> > 
>
>  
>
Received on Sunday, 4 May 2003 15:30:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:42 GMT