Re: Rules Engine [OT]

From: <D_at_B.A>
Date: Wed, 22 Aug 2001 22:23:13 GMT
Message-ID: <lhWg7.11198$2u.81012@www.newsranger.com>

In article <3B841D31.AE726E3D_at_objectarchitects.de>, Wolfgang Keller says...

Quoting your rules paper:
>Deeper analysis yields that your domain experts have trouble expressing their >knowledge in rules. A closer look shows you that writing rules is pretty much >like writing code.

Now after half a century of programming languages evolution, we have rich programming model with exceptions, transactions, variable scope control, threads of execution, and objects. And I fail to identify those constructions in rules programming. What kind of code quality can be expected from the programming model with such a limited means? Yeah, sure, rules would be easily maintainable, just find a knowledgeable rules engineer on the market...

I'm not even talking about rules consistency: what if we have conflicting rules, does the resulting state depend upon the order of execution? Rule priorites? I don't have any time to invest into this bu**it.

BTW, Chris Date have some publications on rules and databases, but it's too simplistic, IMO. Received on Wed Aug 22 2001 - 17:23:13 CDT