- From: Drew McDermott <drew.mcdermott@yale.edu>
- Date: Sun, 14 Nov 2004 20:15:26 +0000 (GMT)
- To: public-sws-ig@w3.org
> [Daniela Claro] > I have been using BPEL4WS for composing web services, and now I've changed > to OWL-S mainly because of semantic features. But how can I use stateful > conversations in OWL-S, Basically, the same way you would do it in BPEL, by having the conversation open with the exchange of correlation tokens that allow each side to keep track of which messages are part of that conversation. Then each process keeps track of its part of the state, associated with the appropriate correlation token. > how can I garantee that I will receive a result in a > long-running business process. I don't understand this part of the question. > And even more, I know that in oWL-S we do not have a mechanism of fault > handler, but there is another mechanism to catch the errors? No, although in my opinion it's a gap worth filling. For the time being, there is no built-in concept of a process being in a "failed" state or having to undo something. The person designing a process might informally label part of it as the failure-handling part, but that label would be just a comment as far as Owl-S or its interpreter are concerned. -- -- Drew McDermott Yale University Computer Science Department
Received on Sunday, 14 November 2004 20:16:01 UTC