Mark Proctor wrote:
We are trying to get Drools into the Magic Quadrant report, however we
cannot do this unless we show that Drools the JBoss Rules engine has
atleast a 3 million "impact" on the market. Because open source is
different we can include end users as calculated financial impact, to
do this we need to gather some answers to questions from our users. No
information will be made public, its purely for us to present to
gartner, and we can sign an NDA if necessary. If you can help please
answer these, as briefly or as detailed as you can manage, and reply to
me directly.
Not sure if my response will help, since my project is not blessed (yet)
by a corporate (ie money spending) entity, but here goes:
1. What specific problem(s) were you trying to solve?
Abstract business rules out of JSP files in a web application, and into
a database.
2. Describe the project itself; what was needed?
Dynamic pages are pre-generated and stored on the webserver docroot. The
pre-generation process is event driven, and pre-generation of a
specified set of corresponding pages is triggered each time the
underlying asset is changed. The mapping of content type to the set of
pages required to be generated is non-trivial (ie spans multiple if
statements, both nested and not).
3. Is this project internal or external and how many estimated users?
Internal, no direct users, 3 content management systems interface with
it, each with approximately 15-20 users.
4. What products & vendors did you consider?
Mandarax, Drools
5. Why JBoss Rules? How much did open source play in your decision?
This is a skunk-works project. I thought it would be a cool idea to get
the rules out of JSPs, so I dont have to deal with the process of
redeploying the entire webapp each time a business rule changes. The
website is constantly evolving, so new pages are added and old ones are
dropped for given content types quite frequently.
FOSS played a big part in the selection. Since I did not know much about
rules engines at the point when I started, I needed something to play
with that would not cost me money. I am ok with none or limited tech
support since I can dig in to the source for things I dont understand.
6. Describe the implementation/size (hardware, OS, databases, # of
servers or CPUs, # of developers, etc...)
hardware: x86_32 (development)
os: linux fc2 (development)
#-developers: 1
7. How long did it take you to implement your JBoss Rules solution?
How did that compare to previous projects?
I am doing this in my spare time, which is about 2 hrs/day, off and on.
This one is my second implementation of the same project. The first one
worked, but I wrote the client directly against the Drools API, and the
result was quite ugly code-wise, so this time I decided to build the
database interface to abstract out some of the Drools calls from the
client code. It took me about 3 weeks to develop the database api and
the web tool, and validate against the "Hello World" and "Petstore"
examples. The next step is to write the rules for my application, which
is not done yet.
8. What benefits do you realize? (ROI, cost savings, performance
improvements, developer productivity, etc...)
Not sure about ROI, cost savings and performance improvements. But this
will help the current maintainer avoid deployments (I am no longer
maintaining the software whose problems this project is designed to
fix), which is a time-waster, so in that sense I guess developer
productivity should increase as a result.
9. Are you now in deployment? How has JBoss performed so far? What
were the main delivered benefits from Drools?
No the project is not complete yet.
10. What is your experience with JBoss support thus far? What support
products do you use (Professional Support, Consulting, Training)?
None.
11. Knowing what you know now, would you choose JBoss again for
upcoming projects?
Depends on need and applicability.
Many Thanks
You are welcome. Many thanks to you and the Drools team for some
excellent software.
Mark
-sujit
