Had a great call last week with a major US Financial Institution on the role, status and benefits of the OMG Decision Model and Notation proposed standard.
What was interesting for me was the Q&A and feedback from the enterprise architects and IT leads of this organisation:
- Should DMN be more that just a decision table standard and also cover orchestrations / ruleflows?
ANS: in time, for sure. Whether this should be added to the scope of a version 1 of a standard is not clear (the DMN RFP allows it for sure). On the other hand, a standard for orchestrating rulesets should use an existing orchestration standard, which in business modelling terms today means either BPMN – an existing standard – or possibly EPC. - Should DMN cover things like rule verification and validation – in other words how will semantics be standardised in real use?
ANS: this was the exact same concern raised by business rules guru Ron Ross in discussions about DMN before the RFP was finalised. It’s true: a standard for things like decision table modelling would also benefit from methodology and best practice support such as how and when to display discrepancies across expressions, such as duplicate rules, for example. However, standardising that behavior is not necessary (again) for a version 1, and is certainly not done in other standards like BPMN. On the other hand, the semantics of business decision models like decision tables will certainly be specified as part of the standard… - Should the vocabulary used in decision models be W3C OWL expressions?
ANS: Possibly; however W3C OWL (2) is still evolving, and there is no business-language version of OWL expressions available. Possibly OMG ODM could be used instead (mapping OWL to UML), but UML objects are not considered “business-readable models” by some. OMG SBVR does provide a business ontology language, but so far the SBVR and OWL camps are not yet working on a join / bridge / resolution here (consider that they are business consultant vs academia led, OMG vs W3C based, and both pre-major-adoption).
The DMN RFP again allows for this evolution by not specifying any particular vocabulary. Some subset of OMG SBVR that maps readily to UML (and executable) objects would seem to be the obvious DMN vocabulary choice for now… - How should non-vendor organisations support or influence the development of DMN?
ANS: Firstly by insisting that their vendors get involved. Secondly by participating directly (e.g. in OMG) with use cases and requirements. The first meetings on DMN development will probably be announced for the next OMG Technical Meeting in UT next month…
The DMN RFP authors would welcome presenting on DMN to other BRMS-using organisations. Let us know…