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

Re: About output derivation in composite processes

From: David Martin <martin@ai.sri.com>
Date: Mon, 28 Apr 2003 07:12:03 -0700
Message-ID: <3EAD36B3.DCCF9B5E@ai.sri.com>
To: Shahram Esmaeilsabzali <sesmaeil@cs.uwaterloo.ca>
CC: www-ws@w3.org

Shahram -

Thanks for raising this question.  It's an area where the DAML-S specs and
examples definitely need additional clarity.

Shahram Esmaeilsabzali wrote:

> Here I quote a comment from
>
> http://www.daml.org/services/daml-s/0.7/CongoProcess.daml
>
> "The following inputs and outputs of FullCongoBuy are derived
> from the corresponding inputs and outputs of its atomic subprocesses
> (as indicated by their sameValues properties), and will normally
> be computed automatically by DAML-S tools.  For example,
> FullCongoBuy uses an input, congoBuyBookName, that's the same as
> the bookName input to atomic process LocateBook.    What follows
> is a partial example of these inputs and outputs."
>
> My question is: how a tool would be able to derive the "output" for
> composite process in an upper layer? the output of a specific atomic process
> could be simply consumed by an input of another atomic process and never be
> seen by the composite process.
>
> Can we use "sameValues" and "Valueof" to resolve the proper outputs from
> atomic processes?

Yes, that is the idea behind the comment quoted above.  If the service developer
specifies the inputs/outputs of atomic processes, and the dataflow between
atomic processes, using sameValues, then a tool can figure out the
inputs/outputs of the upper-layer composite processes.

> if that is the case then there is no "sameValues" and
> "Valueof" in Congoprocess example, how are we going to figure that out in
> this example?

CongoProcess does have some sameValues constructs, in a separate file
(CongoProcessDataFlow.daml).  However, it is true that the example is not
complete in this respect.  Therefore, as you say, a tool would not be able to
deduce the inputs/outputs of the composite processes.

Regards,
David Martin

PS - Sorry about the slow response.
Received on Monday, 28 April 2003 10:12:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:41 GMT