Ernest and all,
I agree with your assessment of that article as it was written,
so let me clarify a couple things. (BTW, I recently used this
paper in developing a proposal for a company I'm currently
consulting with, and recognized the same two observations
as I was re-reading it this time. Great minds think alike.)
First, the ROI assessment was originally written in the context of "Business
Rules Systems" in which you expect not only to train developers
in the rules language, but you expect to train end-users in modifying
the rules as well (which is a whole different ballgame).
Assuming you don't have the latter requirement, but are only training
developers for internal development, the projected ROI is certainly
much less: a month or two at the most.
Secondly, interestingly enough, the application that I proposed using a rules-based
implementation for is precisely a data(base)-driven, indexed-lookup kind of problem,
which an engine like Jess would solve nicely. Although some might argue that indexed-lookup
problems fall under the conceptual umbrella of "significant decision-making capability"
(see question #1 in the article), for practical evaluation purposes, this class of
problems deserves to be mentioned explicitly, I think. (There may be others as well...)
Thanks for the comments!
George
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, April 01, 2003 9:53 AM
To: [EMAIL PROTECTED]
Subject: Re: JESS: Is it for me...
Here's a nice article entitled "Some Guidelines for Deciding Whether
to Use a Rule Engine," written by our friend George Rudolph:
http://herzberg.ca.sandia.gov/jess/guidelines.shtml
This article discusses many of the important considerations. I
personally think it's a little pessimistic; for example, the statement
that the break-even point on ROI comes about a year into the
project. I think you start to see positive ROI as soon as the
developers get comfortable with the rule language -- after a month or
two.
The article also doesn't discuss the efficiency
considerations. Writing a rule-based system with Jess has the
efficiency characteristics of a database-driven application: whereas
for small data sets, there's some overhead compared to writing things
directly in Java, for data of any appreciable size, the indexing,
lookup, and "query-optimization" capabilities of Jess make it much
faster than a naive Java application could ever be.
I think Jeff Richley wrote:
> Ok, forgive the ignorence here, but I do have another
> very basic question. If I have to write the rules in
> the JESS language, what am I really getting over just
> writing them in pure Java? I have to have some big
> selling points to take to my upper ups. I would love
> to use an expert system, please help!
>
---------------------------------------------------------
Ernest Friedman-Hill
Distributed Systems Research Phone: (925) 294-2154
Sandia National Labs FAX: (925) 294-2234
PO Box 969, MS 9012 [EMAIL PROTECTED]
Livermore, CA 94550 http://herzberg.ca.sandia.gov
--------------------------------------------------------------------
To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]'
in the BODY of a message to [EMAIL PROTECTED], NOT to the list
(use your own address!) List problems? Notify [EMAIL PROTECTED]
--------------------------------------------------------------------
