| BPEL or Business Process Execution Language (an | | | | illusion that they can implement a business model with |
| XML format) was created according to the vision that | | | | IT that will allow them to run the business by remote |
| process definitions should be interchangeable between | | | | control. |
| BPM vendors. While that sounds like a noble target, I | | | | For decision making and monitoring the business |
| question its validity. Today BPEL is only supported by | | | | achievements against goals, it is necessary to have |
| that vision as in reality it is unfulfilled. I will explain why | | | | access to real-time business data. BPM people speak |
| that is my point of view and why it causes opposition. | | | | of using ‘a common semantic set’ when they |
| People either subscribe to that vision or they are | | | | mean metadata definitions that correlate to the service |
| implied to have deeper, immoral motives such as | | | | interface data. Therefore not only BPEL has to be |
| locking customers in. The truth is that any product | | | | compatible but also business architecture metadata. |
| (BPM or not) locks a customer in, one way or the | | | | BPEL offers nothing for both needs at all. Presenting |
| other. The most successful vendors in doing this are | | | | real-time business data to the business user in the |
| IBM, Oracle, Microsoft and SAP. So maybe BPEL is | | | | process requires Java code to pull the data from the |
| not a visionary concept but just an ambition for more | | | | service interface and present them. |
| market share. The current expectation is that once | | | | The other element that is needed for pragmatic use of |
| these monopolists will fully endorse BPEL they will push | | | | process is business rules. One can convert business |
| the poor rest out of the market. The means could be | | | | rules into BPEL, but then they cannot be executed as |
| a standards-based process language called BPEL. | | | | complex rule sets that are triggered by data changes |
| There are over 200 other BPM vendors in all | | | | or business events. Business rules also act on |
| imaginable flavors. Less than ten percent of those | | | | business data so you need to code more Java to pull |
| vendors support BPEL and their implementations are | | | | them from the service interface and pass them to the |
| not fully compatible. Therefore it is strange that some | | | | rule engine. Business rules have however multiple |
| claim that BPEL is the ‘de-facto’ industry | | | | purposes with the most important one being so called |
| standard. Yes, it is the only thing that comes close to | | | | boundary rules that are essential for auditing. The |
| something that looks like a standard. Of all the | | | | better your boundary rule set is, the better you will |
| languages that anyone would want to code, BPEL is | | | | catch exceptions and violations automatically. Thus |
| (like any XML format) not something you want to | | | | those rules should not really be an integral part of your |
| write natively. Certainly not a business analyst. But the | | | | process but rather an independent definition but tightly |
| vision and claim is that BPEL resembles process code | | | | integrated into the process execution. There are some |
| that can be taken from one BPM vendor to the other | | | | non-BPEL product that handle real-time data and rules |
| with little to no effort. That, I am sorry to say, is an | | | | quite well. |
| illusion. Where it is part of a sales pitch it is a straight lie. | | | | Some vendors propose BPM 2.0 and other visionaries |
| It is true that you can upload the BPEL code, but that | | | | already propose BPM 3.0. They say that it will not |
| brings hardly a benefit. | | | | happen without BPEL. I think they are right. Therefore |
| When you look at how various BPM environments are | | | | there will be no BPM 2.0. Or even 3.0. BPM in its |
| implemented then you will find each one to be utterly | | | | current rigid form with or without BPEL and BPMN is a |
| unique. The possible combinations of operating | | | | dead-end. New technology concepts are required. |
| systems, Java server versions, database servers, | | | | What would this new technology need to look like? If |
| business rules, content integration, GUI frontend | | | | any technology concept can foster successful |
| functionality, browser and portal functionality, backend | | | | process management, it has to be business facing and |
| application service interfaces (SOA or not) and more | | | | enable real-time metadata-driven model execution, |
| (such as security interfaces) go into the several | | | | while engaging business directly in continuous process |
| thousand. BPEL exists in multiple versions (1.0, 1.1, 2.0 | | | | change cycles. To enable business participation, a |
| are quite incompatible) and allows proprietary | | | | secure change management environment is required |
| extensions most of which are in native Java code. So | | | | that controls the lifecycle and manages deployment. |
| it is an illusion that you can take the BPEL code along | | | | The only way that business users can be involved in |
| and then simply run it in another BPM environment. | | | | designing processes is by using graphical means of |
| Let me take this further. As you do not want to | | | | process representation that has to include rules and |
| maintain BPEL code, why would you even worry | | | | event handling. Non-technical business analysts must |
| about it? What you want to maintain is your | | | | be able to modify processes, rules, and presentation |
| processes and these should be maintained in a | | | | and content. In difference to widespread opinion, |
| graphical manner and if at all be saved in something | | | | process management does not improve business |
| like BPMN, which is another dreadful XML format but | | | | agility by mapping out all activities in flowcharts. That |
| at elast represent a graphical notation. BPEL and Java | | | | was ok when companies worked the same way for |
| means that you need to work in Eclipse, which is a | | | | decades. Today companies need to be able to adapt |
| programming tool (another so called ‘standard’) | | | | on a weekly basis. The best way to plan and |
| and nothing else. BPEL is a functional subset of BPMN | | | | represent a process is by defining the user roles |
| and thus a true roundtrip is not possible. | | | | involved, the data entities required and how they are |
| BPEL does not solve any other issues of BPM like the | | | | serviced by backend applications, business rules, user |
| necessary process monitoring. A customer service | | | | presentation and content definitions. Content state |
| centered organization does not care how you execute | | | | drives the process forward. |
| a process as long as it meets the business goals. So | | | | The problem of complex summary states of content |
| what we need is to monitor business data that | | | | has led to innovative functionality. Consider a software |
| represent those goals. Some of them may be process | | | | agent who will monitor user activity in a business |
| related such as elapsed time for completion of a | | | | process (case) and automatically discover the data |
| customer service task. Monitoring business data | | | | and state patterns that repeatedly cause user activity. |
| requires access to those data and they need to be | | | | Conclusion: When BPEL and even BPMN are looked |
| defined. In what way BPEL would support | | | | at from a pragmatic business need perspective the |
| standards-based tools for business monitoring has not | | | | only acceptable benefit is that products that use them |
| been explained. | | | | have similar functionality and people designing |
| BPM as a principal concept represents also a vision | | | | processes will find it easier to switch products. Given |
| where the noble target replaces pragmatic and realistic | | | | that with other products that low level of technology is |
| achievements. There are no vendor independent | | | | never touched, that potential benefit becomes a moot |
| studies that prove that rigidly managed process flows | | | | point. You can line up BPEL with Java and SOA. |
| are beneficial to a business. Executives chase the | | | | |