Just to record, for clarification, that I have not based my career around any particular product or technology. At this present time in my career, I work for a company that specialises in a particular product set. My career preceedes the advent of these products by about two decades.

For example, Gartner receives consulting” or “seminar” money from the same vendors they are evaluating for their “magic carpet ride”…. !! Vendors wine them and dine them very well. The entire industry is full of “corrupt” practices accepted as the “standard order of business”.

The same is true for all these web sites and trade rags with “writers” and “bloggers” blogging for money, undisclosed, third party press releases… .the entire industry needs to be reformed, in my biased opinion.

Hence, before you can all your software “RETE” you must open the kimono (no licensing muzzle orders) to an independent group of people. The entire software industry is, for the most part, “marketing fraud”, without oversight and controls, but pregnant with conflicts-of-interest, undisclosed relationships, etc.

Of course, there needs to be a well written charter for that organization, with a set of guiding principles, etc. that defines the organization before it seeks membership.

I have seen some strong cross-discipline expertise in the multi-sensor data fusion field, supporting US DOD, where folks seems to collaborate with best-of-breed tools to solve complex multi-agent, multi-discipline architecture, but even then, there does exist professional bias.

However, in the commercial space where the dialog is most driven by commercial vendors, cross-discipline expertise is rare. Most people are pressured to be sales people, not objective system architects recommending various best-of-breed systems and engineering them to work together.

Perhaps we are on the verge of a renaissance in this area as cloud-oriented computing services become more mainstream?

Maybe we should draft a new charter for an organization of people who are not vendor sales driven, not “my company wants me to do this and that driven…” and are truly a federation of objective cross-discipline system architects without conflicts of interest?

When I worked with TIBCO BE, I never considered BE an inferencing engine, because it was not really a “detection engine” as much as a “scheduling engine”. For this reason, I always viewed BE as a type of “core controller” that would be used to schedule complex transactions (which it does very well).

For me, a forward-chaining event-driven rules-processing IDE such as BE was an excellent Blackboard Controller, scheduling and directing the work of many intelligent agents.

So, when I worked on these types of projects before, I considered CEP a larger, more complete architecture, that required numerous components to detection complex events and situations.

The primary reason I placed BE at the center of the BB architecture as the “BB Controller” function was based on the US DOD JDL architecture that is used to process complex events and sensor data.

It make no sense to me to reinvent the wheel, and the JDL is a very good, mature, and quite proven functional architecture for this type of processing.

Unfortunately, I have noticed that TIBCO has (seemingly) dropped this approach, and simplified their marketing material to a version of CEP that looks like a rules-engine; so they have (seemingly) collapsed the vision to match their product v. showing the product as a part of a larger architectural construct.

This is one key problem when vendors are defining the space v. outside experts; i.e. the folks who spent years defining the JDL for the DOD. The JDL is a large scale distribution processing vision that can be applied to a very large class of complex problems.

Reducing the vision to what fits a rules-engine is a bit like burning all the books in a library that are not about rules.

Maybe one issue is that that rule-processing, like RETE, is such a fascinating area, that folks who work deep in that area do not really have the time or energy to go deep into other areas?

http://haleyai.com/wordpress/2008/03/11/goals-and-backward-chaining-using-the-rete-algorithm/

Haley rule engine implemented this approach, so oracle has it now.