vincent
10-28-2008 05:04 AM
So here we are at BRForum08, which kicked off with some tutorials. We missed Sunday�s worthy sessions, including Ron Ross� overview, Neil Raden and James Taylor�s perspective on Decision Management, and Michael Krouze on the value of BREs in SOAs. On Monday there were:
- Gladys Lam on rule discovery (or �harvesting�) -* would be interesting to see how increased in awareness of semantics and automated techniques for rule discovery were impacting the rule analysts� business, if at all.
- Roger Burlton on Process Architecture - which shows how the Enterprise Architect area is beginning to converge with BPM and Business Rules.
- Scott Ambler on Agile Model Driven Development. This was a well-attended session, and Scott is a great evangelist for �agile�. Unfortunately he did not seem to have tailored his message much for a business rule audience, as he was still talking about �programmers and coders� (of which there was exactly 1 in the room). Nothing about how the business rules approach aligns to �agile� in what ways, agile issues in rule projects, etc [*1]. Nor much on how modern tools like TIBCO BusinessEvents provide the model-driven approach out-of-the-box (e.g. as �executable specifications�) - indeed Scott commented more on the perils of round-tripping (whereas in fact you NEVER hack the output code from a rule engine like BusinessEvents - you extend it in the model if you want it to do something different - that�s why its called �model driven engineering�). Agile should be one of the great strengths of CEP technologies: model-driven, business facts defined from business events, rapid execution, built-in persistence, etc. I�m surprised no CEP vendor has contacted Scott to do a White Paper on this� .
- John Rymer talked about analyst Forrester�s framework for selection a Business Rule platform. This looked the same as last year�s tutorial, and similarly made no mention of CEP tools like TIBCO�s rule engine. Even though TIBCO BusinessEvents is probably #3 in the market in terms of BRE sales. Go figure.
- Dr John Elder presented on Data Mining mistakes. Shame I missed this as it looked quite interesting, although of course �event mining� is more interesting .
- Another Dr, this time Dr David Simchi-Levi, presented on optimization technology. I missed this too - it had �sold out�. I noticed though some mention of real-time optimization techniques involving business rules - typically this would be a form of CEP, with one abstract event being �time to reoptimize!�. A great quote, and reason for the increase in awareness in Linear Programing and other techniques, was the net increase in algorithm and hardware performance leading to a 5.3M x improvement in optimization performance from 1988 to 2004. Which is why such techniques are also relevant to real-time CEP.
Notes:
[1] Scott is into audience participation. One of his questions was: what process do you use to build an e-commerce web page with interdependent fields (e.g. zip code and state). The answer is (of course!) to model / design via a rule-driven Form. But that wasn�t one of the options discussed�