- From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
- Date: Sat, 10 May 2003 08:21:04 -0600
- To: public-ws-chor@w3.org
[not wearing my WSA hat, and not reflecting any official position of Software AG] I see basically three different viable options for this WG's mission, ordered by my sense of what is both needed and possible: 1. (probably the best, if other parties agree): Divide the labor with the BPEL TC and work in parallel to define WSDL extensions that would the describe the observable behavior of a multi-part, long-running, stateful set of Web service invocations. This would describe the collective message exchange among interacting Web Services, thus providing a global, message-oriented view of the interactions. Specifically, it would cover the temporal and logical dependencies among the exchanged messages, featuring sequencing rules, correlation, exception handling, and transactions. [Pinched from the WSCI abstract]. The BPEL folks would then focus on the execution language to implement this (and perhaps more functionality). This joint mission would require a lot more collaboration for the greater good of the industry than is probable, but I suspect that the customers / analysts / pundits would think it is the Right Thing. 2. Find the 80/20 point in WSCI, BPEL, WSDL, etc. and build a lightweight, simple, but rigorous (based on Pi calclulus or whatever) choreography language that would be easy enough for most to learn and support irrespective of their BPEL or other stories. Sortof the "RELAX NG" (a simpler, but more rigorous and elegant, alternative to the W3C XML Schema Language) of Choreography. 3. Basically do what WS Aarchitecture should do but doesn't really have the focus or the expertise to do well: Make sense out of the bleedin' nightmare that is WS Choroegraphy / Orchestration / Business Process Execution / Business Process Description / Workflow / etc. Define the principles, describe how they apply to the various "standards" out there now or in production, and generally act as a referee in the food fight that is going on. {Well a couple more, but I don't think they're really viable ... others might disagree on the viability of 4.) 4. Bet on the Semantic Web technologies becoming technologically and commercially viable and do sortof WSCI in RDF/OWL/whatever. The BPEL people wouldn't have to collaborate because it would have the machinery to describe all this stuff at a higher level of abstraction, and tools could use the semantics to generate BPEL, BPSS, XQuery [it's a full programming language, you know], Java, or whatever implementations of the WS-Choreography description. 5. Define the Mother of Choreography Languages (probably some huge XQuery-like thing) that really is both a floor wax and a dessert topping, and builds on and integrates all that has come before. It would take 5 years, but would learn from all the Darwinian processes going on out there now.
Received on Saturday, 10 May 2003 13:45:36 UTC