Jim Sinur

With the recent explosion of the use of rules for agility in many markets, the need for rule management across multiple rule technologies is a growing desire.

When we first asked our clients about having to manage rules across multiple rule technologies/engines, we saw that only about 15% of our rule enabled clients had a need to manage rules across a heterogeneous set of rule technologies. This is changing as rules branch out from specialty rule applications to also include packaged applications, hot spots in home grown applications, business process management technologies, complex event processing, integration technologies and governance focused software. The need for more rule management will grow as rules permeate in a distributed fashion. There are looming issues around rule aggregation, interchange and semantic meaning. Let’s explore a couple of opposing positions for fun

Point: The Rule Interchange Fantasy:

There is no way to manage rules across multiple rule engines and technologies because the rules representations are too different and there is no way to capture the context of a rule. To prove a point, just look at all the rule standards that evolving. There is JSR 94, Rule ML, OWL, ODM, SBVR, RIF and PRR. Each of these are vying to become the common rule language and they all behave like stove pipes without regard to each other. It is a mess at the representation level, much less the context level. If someone expects a common standard embraced by a majority of vendors, they are dreaming.

Counter Point: The Rule Interchange Fact of Life:

Rule interchange is going to happen with or without a common standard. One way would be to provide a transformation hub that would be able to tolerate multiple forms of rules and transform them from one format to another. This approach works today. Another way would be to link a logical business rule to multiple physical rules in a meta-data repository without any transformation, so that common supporting rules could be linked the business rule represented in a linguistic format. The demand will grow and push one or both of these methods until the standards dust settles, if ever.

My Take:

My gut tells me that SBVR has the best chance to be a universal standard because it aims at vocabularies and meanings. Combined with ontology, for context, SBVR would make the ideal compromise to leverage of rules management.