Wednesday August 10, 2005
Don't call it AI The last few days my colleagues and I have been wrangling with various approaches for architecting a provisioning solution. We need to integrate an entitlements system with Sun's Portal Server. The entitlements system will store user/asset mappings while the portal stores user/role mappings. The disparity is that not all entitlements map to a role. An obvious approach is to create a simple webapp for business users to manage those mappings and expose a webservice for the entitlement and portal systems to interact with.
However, considering that I have experienced 8 years of interesting and ever evolving marketing requirements, I wonder if this might not be the perfect opportunity to implement a rule engine. This would give the business the flexibility of developing more complex logic to derive roles from an assortment of entitlements. It would also put the power to change the rules in their hands. The idea behind rule engines is that the business logic can be coded by non-programmers in simple syntaxes - or even GUI abstractions - like drag and drop flow charts. These new rules can be added to the system on the fly like any other data. Essentially the rule engines allow the logic of a system to be changed just as easily as all systems allow the data to be changed.
Looking at rules engines and the Java Rule Engine API, JSR 94, led me to a neat interview with the inventor of Jess (Java Expert System Shell), Dr. Ernest J. Friedman-Hill. I was particularly entertained by this exchange regarding the perception of AI in the job marketplace (emphasis added):
JM: I'm concerned that AI/expert systems experience is still too esoteric for most employers of Java programmers to value as a skill. Am I wrong? How does a Jess developer market him/herself?
EJF: You're right to say that AI experience isn't going to impress many potential employers. But I just did a search at monster.com for business rules and found 1,200 job listings. Like anything else, it's all in the marketing. The cardinal rule of defining AI [is] if it works, it's not AI anymore - it's just programming.
(2005-08-10 07:17:44.0)
Permalink
Comments [1]
I am always glad when I come across someone experimenting with Business Rules Engines.
I should add that business policies/rules are pretty structured intrinsically, but the issue really comes when you need to capture all this in code.
You might be interested in this post, where I discuss how Business Rules can be your answer to complex implementation Clutter!
The other plus, you have added yourself, visibility and control to the business.
Here is a small writeup on some different kinds of business rules that you normally would come across.
Posted by Rajgo on December 27, 2006 at 10:50 AM PST #