Another View on BPEL

Sandy calls attention to an excellent review of BPEL's history and current status from Oracle's Dave Shaffer and Manoj Das in (ironically) WebSphere Journal. Probably the best summary of the differences between the new BPEL 2.0 and the little-lamented BPEL 1.1 standard that I've seen yet in print. She also notes the seeming fakeness of BPEL4People, a joint SAP-IBM white paper that appeared 18 months ago that has achieved what I agree is the highest buzz-to-bang ratio in the history of BPM.

My sources tell me that IBM and SAP have been meeting actively to put forth a BPEL4People spec later this spring, an activity that for some reason the companies' lawyers have shrouded in secrecy. If you recall my original post on this topic, the essence of the BPEL4People white paper is a new BPEL People activity that allows human task management to be integrated more directly with the BPEL logic than is possible with standard Invoke and an external task management service. That means BPEL4People "breaks" BPEL 2.0 engines, except for those with the foresight to implement the People activity. What, you don't have the specs for that? Oh, that's right...

What Sandy also doesn't mention is Devesh Sharma's claim at ARIS Process World that Oracle has submitted its own approach -- standard Invoke plus the external task management service, i.e. compatible with BPEL 2.0 engines -- to the BPEL4People "process" at OASIS. If that's so, it should make for an interesting meeting when IBM and SAP unveil their baby.

Sandy also notes (where does she find the time to read all this stuff?) Dave Chappell's comments on Microsoft's token support for BPEL in Windows Workflow Foundation as a "political" expediency. However, she fails to remark on Dave's real contribution to this discussion, which is that BPMN -- not BPEL -- will be the real portability standard for BPM. I have come to that conclusion myself, but Dave (as usual) beat me to it by 6 months or so.

And then she kind of ruins it by passing on without comment Burton Group's conclusion that "BPEL is just a placeholder for WS-CDL until that choreography standard is ready for prime time." So idiotic I can't even bring myself to click on the link.