Are Process Standards for Real?

I keep hearing the hype around processing standards and I find myself a bit skeptical about the reality versus the hype. While most folks believe that industry standards are a better vehicle for clients than being dependent on several default standards established by the market presence of power vendors, this is a more difficult feat than one would expect. I laud the efforts of the standards groups, but it is an uphill grind that rarely keeps up with the market need.

I believe that business process models should be commonly understood through a visual standard that meets the need of the role of the process professional participating in the creation and execution of a process. I also believe that process models should be interchangeable across platforms and vendors. I believe that processes should be able to execute across multiple process execution engines and it would be best to do that through a common approach.

Where are Standards Right Now?

BPMN is not ideal for business professionals or for technical developers, but it is good enough to gain a common view of a process and its process paths/streams and some detail. I believe that process modelers and process collaborators view their models in what is comfortable to them at the moment. This means a BPMN view and/or a native tool view should be available. BPMN is good enough, but not desirable for most audiences. BPMN does not deal with transport specifications, so the model sharing goal is not quite within the grasp.

XPDL is the most practical transport mechanism available today that can easily support the BPMN model as well as others. I would bet on XPDL as a process model interchange standard, but I am closely watching BPDM to see if it gains traction and is practical enough to use. The combination of BPMN and XPDL is certainly a hopeful combination, but it requires two competing standards groups to work together (OMG for BPMN and WfMC for XPDL). There are a number of industry luminaries like Robert Shapiro, Bruce Silver, and Keith Swenson who are trying to make this a reality. I hope this movement gathers momentum.

UML is quite popular with the developers, but is way over the heads of most process designers. BPMN is the only remnant of UML that makes sense when applied to business process. I view this standard moving deeper into the technical realm.

BPEL is basic in its present state and only handles system activities today, so it is really not what the market ultimately needs. Recent activities have bolstered its capabilities compensating systems transactions to preserve the logical unit of work being completed by a process or a portion of a process. In the real world, processes include human activities and are not necessarily fixed in nature. BPEL is evolving and it will take years before it is truly ripe for anything but system to system process snippets.

Bottom Line:

Process standards are inching along, not as fast as anyone would want, but good enough to visualize and exchange process models. I would heed the advice from Gartner that identified the top five myths of standards.

• Believing something is a standard when it’s not
• Believing a vendor supports it when they don’t
• Believing it’s ready for prime time when it’s late
• Believing everything works together when it doesn’t
• Deciding to wait until everything is final to get started