Re: Rule engines
Date: Tue, 06 Jun 2006 21:35:10 +0100
"Mikito Harakiri" <mikharakiri_nospaum_at_yahoo.com> writes:
> > > What about them? Rule based systems are extremely goofy way to
> > > program business rules. No exceptions, no arithmetics, no
> > > loops. Rule firing order is nondeterministic. I would prefer an
> > > implementation in traditional programming language to rule based
> > > system any day.
> > While your summary is correct for Rete based rules engines, I
> > recently saw a demo by ILog that showed how they have extended the
> > basic algorithm to allow grouping of rules into sets that can be
> > applied in order and based on decision points. It really blurs
> > the distinction between rules engine and programming language.
> > My client is considering using it to allow end users in
> > finance and marketing to change rules without involving the
> > development teams. I'm curious to see how well this experiment
> > works.
[ . . . ]
> Page 6, Figure 2. Business rule which can easily enforced in DBMS.
[ . . . ]
> Page 7, Table 2. This table is designed either by moron or a clever
> salesman who deliberately made it goofy to warrant a case for "rules
Having met them, I'd bet on the latter. They're a French company, after all. ;-)
The appeal for my client isn't that the tool is a better technical solution than SQL or Java, but that the interface allows people from the marketing and finance teams to change the rules without involving a programmer. A lot of thought went into the ease of use of the UI.
In this particular case, the client is using information from Equifax, some industry standard blacklists, and an internal CRM system to decide whether or not to extend credit to a customer and, if so, how much. They're not going to get the rules right initially, so they need to be able to change them quickly and easily.
There are numerous risks with this approach (just try explaining the importance of source code control, let alone rule governance, to a marketeer) of course, but great potential to provide significant business value too. Like I said, I'm curious to see how the experiment turns out.
S P Engineering, Inc. | The experts in large scale distributed OO
| systems design and implementation. pjm_at_spe.com | (C++, Java, Common Lisp, Jini, middleware, SOA)Received on Tue Jun 06 2006 - 22:35:10 CEST