There is a lot of hype and misunderstanding about what BPM is and is not. What you read about and what your peers and managers think BPM is may not be accurate. This webinar addresses this important topic of what “The Four Myths of  BPM Projects” are and what you can do to avoid the mistakes of other IT project teams when implementing process applications. Noted BPM blogger, independent analyst and consultant Sandy Kemsley, of Kemsley Design, and Dr. Michael Rowley, CTO, Active Endpoints, a key architect of industry leading BPM products and ActiveVOS, Active Endpoints BPMS, addressed the following Myths:

1. Business users will create executable process models
2. Business users/analysts can create executable process models
3. Business users/analysts want to create executable process models
4. IT wants business users/ analysts to create executable process models

Sandy explained her conclusions that Model-driven development is a necessity, but not a panacea, and contributions from business and IT are required for complex processes for BPM projects to be successful. Michael provided a demonstration of ActiveVOS, and how easy it is to design and execute a BPM system, all based on open standards. His recommendation was to go through a POC within 30 days, and develop and deploy a prototype within 90 days. This will give the IT project teams the experience and confidence with the ease of use and adoption of ActiveVOS, as well as showing business users the value that can be provided quickly.



Neil-Ward Dutton, Research Director of MWD Advisors, and Dr. Michael Rowley, CTO, Active Endpoints, presented a practical webinar on the theory and practice of BPM and process automation. Neil explained the history of applications development and how the old way of developing requirements and then “throwing them over the wall” is no longer viable. Agile development is what’s needed, and breaking down the barriers between stakeholders – business users, business analysts, developers and operations personnel – must use now Agile development methods to be successful. Michael described how ActiveVOS, Active Endpoints’ BPMS, and its process automation capabilities, easily define the workflow required and immediate runtime using BPMN, BPEL and other web services open standards.

In this episode of CTO Tuesdays, I talked about the similarity between sequence flows (arrows) in a business process and the old GO TO statement from programming languages. In 1968 Edsger W. Dijkstra published a letter titled “Go-to statement considered harmful,” which marked the beginning of a movement within software engineering toward structured programming. Since that time, virtually all programming languages have encouraged a structured development style over the kind of “spaghetti code” that can result from using GO TO statements – until BPM. An executable business process is a program, by any reasonable definition of the word, but the sequence flow of BPMN (the arrow) is essentially a GO TO statement. In this talk, I explained why this is a good thing for business processes, and not the “harmful” thing that Dijkstra saw in the programs of his day.

In this week’s CTO Tuesday’s webcast, I argued (and demonstrated) that any web service that is created using calls to other web services, even when the service being created is entirely synchronous, is best created using a BPEL-based BPMS (preferably with BPMN 2.0 notation). I also briefly talked about the additional advantages that occur when the process isn’t synchronous or when it involves human workflow.

This talk describes BPMN’s concept of boundary events, how they should be used, and how they are related to events that are in the normal sequence flow and to event subprocesses. I also describe the difference between interrupting and non-interrupting boundary events and how processes that use boundary events are mapped to BPEL.

Service-oriented BPM is all about using and providing services. Even tasks done by people are modeled as services. Services use and return XML documents. This means that every decision, every loop condition and generally every use of data has to be able to pull the appropriate data out of XML documents. This is the job of XPath. Many people only have a rudimentary knowledge is XPath, letting their tools generate it for them, but a more complete understanding of the language can help you make simpler processes and allow you a greater understanding is what is going on at runtime. Use the links below to either view a recording of this episode of CTO Tuesdays or just read the slides (the last link).

A number of standards efforts related to BPM are nearing completion of major milestones. This includes 4 standards efforts in 3 different standards development organizations:

• OASIS: BPEL4People 1.1 and WS-HumanTask 1.1
• OMG: BPMN 2.0
• WfMC: XPDL 2.2
• OASIS: SCA 1.1

In this week’s CTO Tuesdays, I describe the current state of each of these efforts along with a brief description of the history and main goals of these standards. Here is a recording of the presentation and a copy of the slides.

In this recording of CTO Tuesdays, I describe the history of the key standards that are important for SOA, such as XML (starting back with SGML), XML Schema, SOAP, WSDL and BPEL. I also describe some of the key architectural characteristics of SOA that drove the standards, as well as some of the standards-making politics that was peculiar to service-oriented standards. If you are curious about the history of SOA and its related standards, you may find this talk to be interesting.

In this recording of CTO Tuesdays, the BPMS podcast, Michael Rowley describes how we got here — taking a special look at previous attempts to solve some core development problems. Whatever your interest: SOA, BPM, application development, even just a passing historical curiosity, you will want to watch this (and future) episodes.

We are very pleased to announce that, once again, ActiveVOS BPMS grew substantially in Q2 2010. BPM users around the world are looking for a new kind of BPMS — one that is easier to master and use. And they are finding it in ActiveVOS. The attached press release has the details of the BPMS’s growing momentum.

OK, OK…I know. The pun on SOAP and soap opera is a little much. But doesn’t the SOA world feel like a never-ending, overwrought daytime TV drama?

I mean, c’mon. Nobody can decide if SOA is dead or alive…if it’s a product (or set of products) or if it’s JBOI (just a bunch of ideas, a pun on “JBOD.” I just can’t help myself.).

So, starting tomorrow on CTO Tuesdays, the BPM podcast, Michael Rowley will begin another “miniseries” within the larger podcast that begins with the very basics of SOA and builds over time to paint a complete picture of this much discussed and often misunderstood development approach. We intend this as a primer for both new and expert users and we are excited that the recurring SOA topics will expand CTO Tuesdays’ regular line-up of BPM technology talks.

Register for CTO Tuesdays at http://www.activevos.com/ctot and, as always, you can return to this blog for replays. But we hope you can join us live because we expect the discussion after Michael’s presentation on these topics to be very lively and we hope to have you join in the discussion.

If you’ve been attending the live recordings of CTO Tuesdays, our BPM podcast, and/or watching the replays, you know that we have stuck to our knitting for the most part: detailed technical discussions of BPM technology.

Starting with last week’s CTO Tuesdays and continuing with the episode posted below, we have increased our range to crucial technical decisions for BPM users which may have very long-term effects.

We are, to put it bluntly, very concerned that the marketplace is receiving — and accepting — incorrect information about the real relationship between BPEL and BPMN 2.0. Last week, Michael Rowley dispelled this myth in the abstract. This week, Michael has gone further: he actually shows what a two-toolset, two-engine BPMS environment with only a fig-leaf of integration looks like, using Oracle’s BPM Suite 11g and SOA Suite 11g as the poster children.

Yes, Oracle is a competitor. And yes, we have a “dog in the hunt,” as they say. Therefore, for sure, we have an opinion.

None of that undoes the fact that users should consider alternative points of view — views based, as we attempt to do, on the exact text and meaning of the BPMN 2.0 specification. And the fact that we have an opinion — and a product based on that belief — doesn’t undo the fact that much of the argument that BPMN should execute directly and that BPEL is passe is as self-serving as anything we may say.

So, I urge you to watch the replay of CTO Tuesdays attached to this post and to consider the alternative arguments we make. We’re not going to convince everybody, but we truly believe that the people who do consider their long-term BPMS strategy will find that BPMN as notation with BPEL execution is the better alternative.

If you follow Active Endpoints and ActiveVOS on the web and/or in social media, you know we aren’t timid about…well…anything. We try hard, however, to make sure that as we forcefully make our points we are backing them up not just with emotion (a remarkably clarifying attribute often missing vendors’ discussions of technology) but also with hard facts.

You can see the very best of that loud-but-authentic aesthetic at work in episode 28 of CTO Tuesdays, the BPM podcast.

For too long — and, frankly, for reasons that mystify me — some voices have spoken of a “conflict” between BPMN and BPEL. It has always seemed to me that those points of view — that BPMN 2.0 is somehow a “successor” or “replacment” for BPEL –  have an agenda that’s more about their preferred results in the marketplace than about the “best” or the “right” thing for users. IOW, politicking is at play.

But, as we all know, in politics, negative campaigning works. “BPEL is dead;” “BPMN 2.0 execution obviates BPEL.” These misstatements have gained far more attention than they deserve. They have escalated to the level of myth — or worse, conventional wisdom — both of which can have lives very separate from reality.

Our response: BPMN 2.0 is better with BPEL execution for users for a plethora of reasons. Far from dead, BPEL’s fundamental mistake of not specifying a visual notation is cured by BPMN 2.0. And BPMN 2.0 achieves its highest likelihood of success when coupled with BPEL execution.

Still, the myth that these two crucial standards are in conflict persists. Watch this podcast replay to see and hear Michael Rowley debunk these myths — passionately and accurately.

And be sure to join us next week for episode 29, titled “Oracle’s misguided approach to BPMN and BPEL” for even more myth-busting. Register at http://www.activevos.com/ctot.

Every once in a while I find someone repeating the common myth that BPMN and BPEL are in conflict – that you have to choose one or the other. The most recent place I saw this was in Tom Baeyens’ rebuttal to my criticism of his microkernel-like approach to BPM engine development for Activiti. In that article he references an article by William Vambenepe that shows a completely invalid example of a conflict. I will show the mistake made there, but before I do I’d like to make a more important point on this subject:

BPMN 2.0 Complete Conformance can only be claimed by an engine if the engine also supports the BPEL Process Execution Semantics Conformance Type.

The phrases in bold are the names of conformance classes in the latest public draft of the BPMN 2.0 standard specification. The conformance section of the specification defines multiple conformance types; one of which is the “BPEL Process Execution Semantics Conformance Type”, which defines how to use BPEL to execute a standard BPMN model. The last BPMN conformance type is called “complete conformance” and it also requires support for BPEL.

So, back to the article that Tom Baeyens’ linked to when he claimed that that “the translation step from BPMN to BPEL is very problematic to say the least.”

As it turns out, William Vambenepe misunderstood the semantics of the BPMN construct that was supposedly in conflict with BPEL. He references this snippet of BPMN:

But he describes it this way:

The customer quote can be reviewed by the region manager, the country manager or the VP of sales. At least one of them must review the quote. More than one may review the quote.

He then goes on the show how hard it is to represent the at-least-one requirement in BPEL. The problem is, the above BPMN snippet has no at-least-one semantic.

Here is what the BPMN 2.0 specification says about the inclusive gateway: “each path is considered to be independent, all combinations of the paths may be taken, from zero to all. However, it should be designed so that at least one path is taken.”

This means it has the exact same semantics as BPEL’s concept of conditional links out of an activity: any subset can be followed, including none.

People also sometimes claim that the problem comes from the fact that BPMN is unstructured while BPEL is structured. Actually, the problem is that some tools don’t know about the free-form style that is permitted in BPEL. BPEL supports both structured constructs and unstructured flows. Oracle BPEL Process Manager, for example, does not show the links in unstructured flows (no arrows), so they are basically worthless in that tool, but the standard does allow them and ActiveVOS supports them fully.

So, are there any processes that can be represented in BPMN that are difficult or impossible to map to BPMN? Yes, there is a restriction in BPEL against cycles in flows that make it difficult to represent interleaved loops in standard BPEL (although I haven’t actually seen an example of this pattern in a post about the mismatch between the two languages). However, this restriction in BPEL is not is fundamental to the language. Active Endpoints has implemented this simple extension that removes that restriction and we encourage all BPEL engines to also support the elimination of that restriction.

Let me finish by quoting the first two paragraphs of the BPMN 2.0 specification. Note especially the second paragraph.

1. Scope

The Object Management Group (OMG) has developed a standard Business Process Modeling Notation (BPMN). The primary goal of BPMN is to provide a notation that is readily understandable by all business users, from the business analysts that create the initial drafts of the processes, to the technical developers responsible for implementing the technology that will perform those processes, and finally, to the business people who will manage and monitor those processes. Thus, BPMN creates a standardized bridge for the gap between the business process design and process implementation.

Another goal, but no less important is to ensure that XML languages designed for the execution of business processes, such as WSBPEL (Web Services Business Process Execution Language), can be visualized with a business-oriented notation.

Clearly the specification writers see no conflict between BPMN and BPEL.

In the loosely-coupled world of today’s process applications, how does the BPMS know which running process instance to route incoming messages to? It’s an important question…and one CTO Tuesdays has explored before (see CTOT #5 on engine-managed correlation).

In this episode, Active Endpoints CTO Michael Rowley tackles a more complex case for correlation: when you can’t use engine-managed correlation because the developer doesn’t have control of the client in a business process. This episode reviews the terms used in correlation and then walks through a simplified procurement process to illustrate the concept.

As always, we welcome your feedback and suggestions. And we hope you will be able to join us for the live recording of CTO Tuesdays ever Tuesday at noon ET.