Eliminate the Middlemen of Business Rules: Thinking Fresh for the New Year and BeyondI have said many times programmers, even semantic programmers, are not the wave of the future for business rules. The future lies with enabling business people and business analysts to have dialogs with machines coming as close to unambiguous statements as they can (e.g., in RuleSpeak). The machine should ask questions to reduce ambiguity to an acceptable minimum. To protect against liability, the machine should log all assumptions, major and minor, in translating to an internal (formal and unambiguous) form so that results are traceable and improvable. Consider IBM Watson. If machines can win at Jeopardy against the best players in the world, that’s a pretty impressive feat for natural language capability. Let’s compare apples to apples. There will always be translation of business ”requirements” from human form to machine form. Even some perfect (a.k.a. formal semantic) implementation language would not reduce translation errors to an absolute minimum. Coders would still have to translate, and errors compound with every translation. So I say disintermediate; eliminate the middlemen – i.e. the coders. If you look across a great many industries, that’s the trend. Why should development of business applications be any different? I also suspect that any “formal semantic” language for business rules would inevitably be English-biased. That’s simply not acceptable in a global economy. RuleSpeak has gathered increasing attention around the world. There are now versions in the works for Norwegian, Polish and Japanese, in addition to the original English and the existing Dutch, German and Spanish translations. A growing number of people find that RuleSpeak strikes just about the right balance between structured and unstructured expression. I’m not saying, however, that other useful approaches couldn’t be more formal – or less formal – than RuleSpeak. Perhaps they could. And that’s been my point for all these years working on SBVR as a business rule standard. We simply don’t yet know the absolute best approach for expressing all business rules in all circumstances. No one knows enough. Perhaps there isn’t one. SBVR is brilliant precisely because it captures semantics without dictating expression form. Long term, that’s exactly what the industry needs. No, there hasn’t been rapid vendor implementation of SBVR since release of 1.0. I wouldn’t expect there to be. It threatens virtually every interface and mindset on the planet. Most people in the IT field still just don’t get it. In retrospect, OMG may have been the wrong forum for SBVR for at least two reasons:
1. OMG is the bastion of best-of-breed programming standards. Obviously there is an important role for that, but as above SBVR isn’t about programming.
2. The SBVR vocabulary is extremely useful for organizing business conversations about business vocabulary and business rules. OMG doesn’t ‘do’ business-facing standards of that kind (i.e., ones that don’t “compute”). IMO, that’s a major shortcoming. There is ultimately nothing more important than improving communication at the level of people. Get it wrong there and I promise it will be wrong in business automation, no matter how elegant the implementation language.The bottom line is that machines for business (rule) automation now must learn to ‘speak’ human languages. The other way around is simply no longer acceptable – or even necessary. www.BRSolutions.com