The 🛣️roadblocks to adopting Decision Model and Notation or DMN.
As I discussed in my last post (https://lnkd.in/gbvVJU-w), DMN is an excellent choice for modeling business requirements around business decisions and rules into an XML-based model that can be ingested by software and potentially executed.
🔶The first challenge and a major roadblock to adoption is the potentiality of the executable model. In DMN, there are three levels of conformance, meaning the vendor platform that claims they are DMN compatible is in one of these levels; note that compliance at one level must also be compliant with any preceding levels:
– Level 1: Visualization of Decision Requirements Diagram (DRD) and Decision Table for #businessrules. Imagine a standard way of illustrating decisions and rules.
Level 2: Partial section of fully executable FEEL, decisions, and rules.
Level 3: Full set of Boxed Expressions and support for all FEEL expressions.
Unfortunately, not many vendors in the space have attempted to become conformance level 3. This optionality for conformance has shot DMN in the foot. It turned out to be an option for cheating the DMN rather than helping toward its incremental adoption.
🔶A big part of the execution of decisions and #rules is pre-execution and post-execution work, which requires data connectivity, modeling, and on-the-fly data processing (transformation). As with any standard, the DMN had to draw the line between what is included and what is not included in the standard. In this case, the DMN left data part of rules and decisions execution out. Of course, there are other standards in OMG that cover these parts, but this approach does not help DMN's strength. In my opinion, it should have been an all-in-one standard that includes at least many of the common scenarios regarding data requirements.
🔶Last but not least, overselling the Decision Model and Notation (DMN) as other things such as #decisionintelliegnce, #decsioncentric from vendors that adopt DMN to some extent hurts DMN more, as sooner or later it will become obvious that DMN is just the standard for decision and rules modeling not a Open Decision Intelligence Platform (DIP). The DIP requires significantly more capability in areas such as orchestration, advanced data processing, stateful long-running decisions, and so on. This misrepresentation devalues DMN rather than helping to adapt for what it is and its purpose.
✨DMN is not perfect! There are many design flaws in its XML, FEEL looks outdated for many areas and use cases and there are countless issues in standard documentation and inconsistencies all over examples and so on. But all in all, I believe it is a positive step forward toward standardization of business decisions and rules that I hope it gets more attention and true adoption at some point in near future.
Post here.
Published May 20th, 2024 at 07:30 am
