Re: XP and non-anectdotal evidence?

Steve Howell wrote:

> If you're comparing two software methodologies in a controlled study, it
> seems like the most important thing to control is the requirements, and
> you'd want to do that up front.  But XP, rightly I think, says that you
> should never fully specify the requirements up front.  Instead, the
> requirements should evolve as the customers (and developers) learn more
> about the software.  Since a non-evolutionary methodology and XP will
> never produce identical software products, you end up having to compare
> things like customer satisfaction, which gets subjective and easily
> tainted by the agendas of the people conducting the study.

Give both studies a pool of end-users, and see if they increase their
productivity.

Now the planned system must lose. The agile system will simply climb the
hill of greatest local productivity gains.

So, even if at the end of the project the planned system happened to achieve
the same productivity heights, the integral or sum-total of the
productiveness would be greater for the agile project. The end-users got the
biggest productivity boost earliest.

By this thought experiment, the only way to fix the planned project is to
make the planning highly responsive and iterative: Agile.

--
  Phlip


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Rent DVDs from home.
Over 14,500 titles. Free Shipping
& No Late Fees. Try Netflix for FREE!
http://us.click.yahoo.com/ybSovB/hP.FAA/3jkFAA/NhFolB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
[email protected]

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/