BPMN and BPMS in BPM 2.0 are the realization of “Application Development without Programmers” as envisioned by Dr. James Martin more than 20 years ago. Workflow engine was a natural extension of Xerox’s research and experimental workflow system entitled “Star” from which GUI and object-oriented programming (Smalltalk) originated.

The ‘we and them’ mindset is a die-hard species! In the early 50’s of 20th century, Rear Admiral, Dr. Grace Hopper (mother of COBOL) punched tapes to program the first-generation computer to compute entries for artillery ordinance tables. Nowadays, five years old children use computers to play and learn. Shouldn’t we erase the artificial line that separates an organization into business and IT communities? If we do, we don’t need any “strategic alignment”!

Is there a big difference between BPMS and Workflow Engine from a business perspective? They are just two sides of the same coin. The one with the most extensibility, under the hood, for “plug and play” software and hardware components will win!

As BPM evolves, it is very likely that it will replace custom designed and built SCADA (Supervisory Control And Data Acquisition) system for real-time process control applications. Examples are production control, intelligent building, area traffic management, remote sensing, patient vital signs monitory instrumentation, and telecommunication network management.

BP (Business Process) is more than order fulfillment for the management of supply and demand chains. There is “information pollution” for many top executives who need in time and relevant data & information, which internal systems can never supply. They need external business intelligent data and information to make strategic decisions. In the 70’s, Operations Research community was asked by oil executives, “Why couldn’t you forecast the oil embargo by the OPEC?”

UML has been a “de jure” standard, adopted by ISO, for extensible modelling language. One can find UML extensions in GIS (Geographic Information System), SysML for systems engineering, and STEP (another ISO standard) in product data description and exchange. Incorporating BPMN into UML to replace “Activity Diagram” is a real blessing!

To the BPMS vendor, workflow is just a pattern deployed on the BPMS. That is, a BPMS can morph to be a workflow engine, just by adding the required process definition (in BPEL) that executes the way a workflow engine does. It can also morph to look like ERP, an SMTP server or any other kind of process. And of course, with swimlanes (utterly separate processes interacting only by communicating), it can combine all these processes into end to end distributed, concurrent, persistent, transactional or collaborative processes. Workflow products cannot do that. A workflow engine cannot be anything other than what it is, a workflow engine. It can execute any workflow, but it can only execute workflows.

Now, all that may be academic to some users of course. I expect BPMS to solve some big challenges that a WFMS cannot. I expect BPEL to solve some big challenges that XPDL cannot. The two paradigms may persist for quite a while. Here is a prediction for 2006:

We’ll see the first next generation workflow systems built on a BPMS. Why? Take a guess.

The larger issue is how Ismael’s BPM 2.0 differs from earlier conceptions of BPM, such as my original BPM 2.0 formulation (”process without programming”) based on my (perhaps faulty) understanding of Howard’s book BPM: The Third Wave. The promise then was that BPM, based on SOA standards, would give “the business” the tools to directly implement executable processes and change them quickly. In reality, however, the complexity of SOA has firmly established IT in the driver’s seat of BPEL implementations, and they don’t want “the business” anywhere close to implementing anything directly. I think this is why BPMS vendors that really do try to give implementation directly to “the business” — like Phil Gilbert — have such a visceral aversion to BPEL.

Ismael’s reformulation addresses this. In the new BPM 2.0, process implementations will done by IT “process analysts”, but require little or no code and only widely available skillsets, not J2EE programming. This is also consistent with IBM’s description of the target user of WebSphere Integration Developer, it’s BPEL tool. In response to some comments, Ismael has added a role for the business in process modeling and simulation analysis as a precursor to BPEL design. In this way, the “new” BPM 2.0 addresses the real elephant in the room, which is business-IT alignment and finding new ways to react quickly.

Quote from the minutes of the joint BPMI-OMG Meeting (Washington, D.C.– November 2, 2004): “The WfMC has not developed a notation for XPDL. They may adopt BPMN. WfMC members are working with the BPMI Notation Working Group to build a mapping from BPMN to XPDL. WfMC sees XPDL as the transport mechanism (XML Schema) for BPMN.”

So there seems to be at least XPDL as a serious contender to BPEL. My impression is that BPMN is a de-facto standard and BPEL is one — perhaps even the most widely used one — ‘transport mechanism’ for BPMN, but not the only one. Also, as Bruce Silver has pointed out elsewhere, there is currently more XPDL than BPEL support among BPMS vendors.

BPEL is the standard for executing processes. Why fight against it?

The elephant in the BPM room is this: BPEL and BPMN won’t provide seamless round-trip engineering; BPEL is a wrong-headed distraction. “Please papa can I go Down to Richmond to the traveling show Please papa don’t say I can’t…

Bruce, I do however agree with you that the definition of BPMS needs to be adjusted. BPM 2.0 as term, misses the point that a BPMS really is a quite radical architecture and computing paradigm. So the real issue is to define this new category more formally. It will be based around a virtual machine for concurrent execution and a design driven architecture (DDA) built around “process” as a new abstract computing entity (subsuming and syntheziing today’s separate paradigms of code, data, messages, rules, etc. I think the reality is, BPM 2.0 is really about the BPMS and what it really is, as opposed to incumbant distortions from existing product categories.