Team,

Before we decide to put both codebases under the same project or under separate projects, I would like to ask a couple of clarifying questions.

First, what is the goal of the Ode project? Is it to develop an implementation of the BPEL specification, or is it to develop a process engine that is based on BPEL, but could support other workflow patterns as well, as was suggested in a previous post? While both options have their own merits, I am a little bit worried that the later could lead to a lot of confusion and wasted efforts, especially if we were to start from two separate codebases that have very different architectures underneath. I have been in the process game long enough to know that not specifying the process model upfront -- be it BPEL, BPML, XPDL or WSFL -- can take one down a path leading nowhere. A lot of work went into the development of the BPEL specification, and I respectfully suggest that we do not try to re-invent this work here, at least not until we get a working version of the process model we pick as a reference point.

Second, if we agree that the goal is to develop an implementation of the BPEL specification, I would like us to answer the following questions. I understand that some questions have already been answered on this mailing list, but summarizing the answers would certainly help.

- Do we want to support BPEL 2.0, BPEL 1.1, or both? - Do we want to be independent from any specific ESB? - Do we want to be independent from any specific J2EE application server? - Do we want to rely on JBI and how does it impact the previous goal? - Do we want to have something that we can release now or can we wait 6 to 12 months? - If we want to release something now, is it ok to have two separate codebases under the same project?