There is a new book on BPMN modelling called “BPMN Modeling and Reference Guide” by Stephen A White and Derek Miers. It was launched at the Gartner BPM Summit event in Washington DC last week.
Net Take Away: This is a great resource for those coming up to speed on BPMN. It uses a lot of practical examples of process diagram, starting from simple ones and working toward the more complex ones. It gives you a good base for learning and using the essential BPMN concepts, while avoiding many of the more esoteric and less useful details.
I particularly like chapter 4, called simply “Modeling Issues”, which covers the topic of universality of models. Here it explains why you are not going to be able to draw a diagram without knowing where and how the diagram will be used. The person drawing the model needs to make a lot of choices on what to include and what not to include in the diagram in order to tell the particular story to a particular audience. I touched on this last year in my post on “BPMN & Methodology Agnosticism” and it is good to see a treatment of this as a caution before jumping into drawing models. It is important to dispel the myth that there is a single “correct” model. Models are simply more or less useful for a particular purpose.
The book takes a middle ground giving examples sometimes for human processes, and sometimes for system to system processes. BPMN was designed for both, so it is reasonable to give examples from both.
No book is perfect. The appendix on execution environment seems to have been time-transported out of last century. It implies that BPML is a good representation of a BPMN diagram — BPML is a language that was put to rest so long ago, bringing it up here is only a distraction — and finally concluding that BPEL is evidence that a BPM Suite “can directly execute a process model”. Everyone living outside of a cave has noticed that BPMN has to be fundamentally transformed in order to produce BPEL, so to say you are executing the model directly is very misleading. In fact, recent research by Jan Mendling (Queensland University of Technology) & Gero Decker (HPI Potsdam) showed that BPMN and BPEL do not support the same basic patterns so translation between them is always going to be problematic. I suspect this appendix was cut-and-paste from some old BPMI marketing literature.
The reader should understand that ability to draw and directly execute models is nothing new: many systems have been able to do this since the early 1990′s. What BPMN brings is a standard dictionary of graphic images for visualizing these models. Too bad this book does not at least touch on some of the problems of creating a repository for sharing and exchanging of process models, but then that is outside the scope of training users to draw BPMN models. So don’t read this book for advice on execution or exchange of models, but instead you will find here an excellent introduction to the subject of BPMN Modeling.