vincent
11-11-2008 02:53 PM
This week sees the BPM Think Tank move to Europe, and a chance to observe any differences in BPM expectations and takeup over the previous US event. TIBCO were here to host the round table on “Combining Event Processing with BPMS for Agility” - more on that later.

As it happened I missed the first few minutes of Jan Baan’s* (of ERP fame) keynote, but walked in just as Jan was presenting on the need for complex event processing, and the need for BPM to handle complex events. Of course, BPM systems can consume complex events just like any “normal” events - its just that detection of complex events requires a continuous event-pattern-recognition engine.

Some of Jan’s other noteworthy points were:

• “BPM 2.0″ is human AND integration centric, with the observation that only 8% of processes today are automated.
  • No disagreement from us there: TIBCO iProcess and BusinessWorks are a common combination in TIBCO’s “BPM+”.
• “BPM 2.0″ includes SOA, RIA, business rules, BAM, model-driven, app-server-free.
  • More “BPM+” endorsement.
• “BPMN has no room for intelligence”.
  • Although Jan was probably talking about human collaboration (which AFAIK BPMN does not prevent), this could be considered true if you consider “intelligent processes” to be where rules drive the selection of processes. Of course, a BPMN diagram is as “intelligent” as the process flow it represents - nothing more and nothing less. This is not a fault of BPMN, just of the widespread (and incorrect) view that “process orchestration is the only way to model processes”.
• “ERP is sh** everywhere, and cannot be replaced”.
  • 
• * “BPM is the replacement for ERP”
  • Presumably this was meant to be considered separately from the prior statement on ERP…

As usual, the 2 Event Processing round tables were interesting, with the first as a diverse audience discussion, and the 2nd discussing dynamic / multi-level BPM with an existing iProcess user. Points covered in the former included:

• The role of events to define interfaces between process parts to allow easier re-factoring.
  • This is an interesting idea, and may be covered by the “process meta metamodel” i.e. BPDM. But I’ll have to check …
• Events and complex events driving ad hoc or unordered processes or case management.
  • Yes.
• Events originating from BPM systems (and being collated with other events) to provide insight into processes or BAM.
  • Yes.
• Events and context driving process exceptions without having to define such exceptions in (i.e. avoid polluting the business view of) the main process flow diagram.
  • Makes sense as a BPMN 3.0 suggestion!
• Whether events can play a role in mapping SBVR rule statements to BPMN process diagrams.
  • Possibly, but note that SBVR is about policy-type business rules, and processes are about business operations, there is no direct mapping from SBVR to process anyway. And if you were to create a process from SBVR statements, it’s probably easier to map to a set of event-driven inference rules rather than an arbitrary process flow. Probably we need some PhDs to check this theory out through the TIBCO University Program!

Source...