Book review: "The Decision Model"
Jean Pommier 270001XBPR firstname.lastname@example.org |
0 Comments | 1,990 Visits
First, hats off to the authors, Barbara von Halle and Larry Goldberg, for producing more than 500 pages on this topic. When you are heads down in client projects all the time, it is difficult enough to find the time to read what is out there let alone contribute in any meaningful way. So, I’m always amazed, and frequently grateful, when others in the field of decision management, like consultants, have the time to lend their perspective and expertise in such detail. I particularly liked the way that Barbara and Larry included the views of other reputable experts throughout the book.
A couple of years ago, one of our key clients issued an RFP which referenced KPI's (Knowledge Partners International) proprietary framework to model and document decisions. I’m glad that the public can now freely access the associated concepts, thanks to this book, The Decision Model, released in 2010.
What I liked the most about the book? Apart from the interesting expert commentaries in Section III which summarize several other books and publications and are worth keeping handy, the most useful content (for me) is in the first section – the discussion that there’s more than data and process in life, that decision modeling is an art and science of its own, complementing other models and modeling paradigms.
For years, if not decades, I feel our modeling views have been limited to two dimensions. On the one hand, there’s the data or information modeling school, dealing with the “what” that any business manipulates. On the other, we have process modeling, focusing on the “how” through the sequencing of tasks. Like business modeling is a key component of BPM, same for rule modeling and Business Rule Management, it is time to promote decision modeling as a key component of Decision Management and The Decision Model leads the way in that direction. With several hundred pages, the topic is certainly addressed in detail, including a scientific rigor, almost a formal proof, which is not going to bring much to most practitioners (Section II). Even with 500 pages on the topic, there is still more to be written on what decision modeling can bring to our world -- the why and the how – so that we can all make better and good decisions.
What I didn’t like as much? The overall limitation of the scope of decision modeling to rule-suited decisions despite a title which conveys a much more generic idea, the entire discussion and focus is on decisions which can be depicted as a collection of business rules. While I’m of course a big fan of business rules after more than 25 years spent building rule-based decision-support systems, and see the value for Business Rule Management, there is much more to decision modeling than just facts, operators and rule families. Some decisions can certainly benefit from automation (event processing, rule management), while others are improved through better orchestration (process management), better instrumentation (business and predicitve analytics, information management, collaboration, advanced visualization), or other optimization techniques. I’m therefore disappointed that the “Decision Model” concept is taken, not only as a book title but also as a patent, in such a narrow capacity. Indeed, like we can model data or processes outside of an IT-driven initiative, for the sake of knowing, identifying and planning areas of agility and performance improvement, decisions are taken everywhere in organizations, and we need to be able to model all of them, not just the ones which correspond to rules.
On this topic of third modeling dimension, I would like to point out that in his Génie Cognitif published in 1988 (available only in French, sorry…), semiotician Claude Vogel was already highlighting three themes to model knowledge and expertise: taxinomies or classifications (we’d say ontologies nowadays), actinomies or actions (or processes…), and interpretations (e.g. inference, deduction, planning). While inference and deduction are well supported by business rules or business events in many cases (and Claude’s book was written in the context of Expert Systems), there are situations, or decisions, which are too intuitive, or not declarative enough to be relevant to rule technology or maybe to any technology at all, yet. To claim generality, we need “the decision model” to include them, therefore a decision model broader than the one defined in this book. Thankfully, there are more than procedural decisions out there, we, humans, have a future beyond Watson as we can read in this article by Seth Borenstein and Jordan Robertson, from the Associated Press: "The way to think about this is: Can Watson decide to create Watson?" said Pradeep Khosla, dean of engineering at Carnegie Mellon University in Pittsburgh. "We are far from there. Our ability to create is what allows us to discover and create new knowledge and technology." Overall? Again, an interesting and solid venture into decision modeling, albeit with a predominant and limiting rule-based flavor. If you are in a rush, you can jump to pages 309-313 for the definition of the model. Or read a public primer published by Jacob Feldman’s OpenRules.
PS – Speaking of decision modeling, here’s an event that you may be interested in if you happen to be in the D.C. area: a special Decision Modeling Information Day organized by OMG around the Decision Model Notation (DMN) standardization effort. March 23, 2011 in Arlington, VA. Hope the group will expand the horizons of decision modeling beyond the decision table-like paradigm!