My latest column on BPM Institute covers the issue of end-to-end process models which involve multiple pools and deeply nested hierarchies. It was motivated by painful experience grading certification exercises in my BPMN training. That experience was critical to my ultimate decision to include this in version 3.0 of my BPMN training, overcoming an initial concern about scaring off would-be process modelers. But the difference between the simple exercises modeled inline in the training and the post-training mail-in certification exercises representing real-world processes was simply too great to ignore. I’m not going to summarize the column here, but I recommend it to anyone interested in BPMN modeling.
For the blog I just want to add that topics like this are emblematic of the evolution of my approach to teaching BPMN from my initial efforts a year ago to version 3.0, which launches tonight on bpmessentials.com. When I started, my goal was to teach the semantics of the shapes and symbols as defined by the spec, especially where the spec does a particularly poor job of it (plenty of opportunities there!). But the experience of interacting with students showed me there is a hunger out there for a methodology, not an analytical framework like Lean (although that exists, too) but more of a cookbook approach representing best practice for BPMN. The spec doesn’t give you that. As a vendor-independent standard, BPMN has no built-in methodology, and BPMN tools vary in their ability to model complex end-to-end processes in the top-down hierarchical fashion that I think is vital. Naturally, the tool we use in the training, Process Modeler for Visio from ITP Commerce, does that part very well, but it still demands a cookbook methodology for organizing models stretching over many many sheets in Visio, so that they not only are easily traceable visually but hang together as a single end-to-end model for simulation and ultimately execution.
I think this article has hit the problem of large process model traversal and modification right on the head. Present BPM toolsets need to incorporate some of the advances in large graph traversal and manipulation methods available from other data visualisation systems, to ease the pain for people dealing with such large models.
Thank you for the article, will be using it as a case study on such issues in my research within the QUT BPM research cluster.
Ross – bpmve.blogspot.com
Now that I’ve had a chance to read the article in full, I’m looking forward to going through the new version of your course.