[webproducers] Re: formal testing benefits

Duane -

Anders covered the user-centered perspective well. Here's another
angle--quality/defects: the earlier and more often you test, the less
expensive (financially, physically, manpower, pick your hot-button) it
is to fix an error. The cost of finding and/or correcting an error (or a
requirement creep) is exponentially more expensive to fix as you get
closer to (or pass) the project completion (some books say up to 1000
times more expensive than catching it earlier...)

Say your team discovers a significant error in the latter half of the
project: at the very least it'll probably take more money or resources
to fix. But you also risk missing a launch date. How about a bug that
makes it to market? Is the PM willing to risk the financial impact of a
product recall? How about company reputation? I know the behemoths like
MS can get away with it ("Oh, you just need to download the patch for
build 12432a!") but not all companies are that lucky.

Between the user-centered and quality-centered perspectives you've
probably got good ammo. User-centered arguments will appeal to the
sponsors and marketing people who don't want to risk a _product_
failure; the quality perspective will typically ring the bells of
developers and PM's who don't want to risk _project_ failure. Either
way, it appeals to the "you're only as good as your last at-bat"
mentality.

(Some references: "Rational Unified Process-an Introduction", Krutchen;
"Rapid Development," McConnell, *especially* ch. 3 & 4)

Greg Glaser
Project Manager
gglaser@xxxxxxxxx
 
 


__________________________________________________________________________
To unsubscribe send a blank message with unsubscribe in the subject to 
webproducers-request@xxxxxxxxxxxxx

To access our webform (instead of sending e-mail) for popular commands 
including subscribe, unsubscribe, digest, and vacation  visit  
www.WebProducers.org. You can also access the list archive at the website.

Questions and comments are welcome just e-mail me, morry@xxxxxxxxxxxxxxxx

Other related posts: