Another academic "study" yields obvious (and flawed) conclusion . . .

Agile Teams Favor Agile Methodology!

Conrad Weisert
April 13, 2012

NOTE: This article may be reproduced and circulated freely, as long as the copyright credit is included.

"In the mid-1990s, the prescribed means of keeping software development projects out of trouble and on schedule was to follow a heavyweight software development methodology consisting of a complete requirements document, including architecture and design followed by coding and testing based on a thorough test plan."

Lead sentence in "What Agile Teams Think of Agile Principles", by Laurie Williams, Communications of the ACM, April, 2012, pp. 71-76

Academic "studies"

We're all acquainted with those quasi-scholarly papers in which the author cites convincing results of a survey to confirm a conclusion that was obvious to everyone from the start. I once sat through a 75-minute talk by an earnest young woman who cited an avalanche of survey data to show her audience that

"User satisfaction is strongly correlated with quality of service."

No kidding!

Now we have six pages in the prestigious Communications of the ACM to persuade the reader that programmers who have been practicing the so-called agile methods for years actually like and enthusiastically endorse agile methods. I was willing to accept that claim before reading the article. But in this case, the whole study rests upon the faulty premise stated above.

Confusing analysis and design

A complete requirements document most certainly does not include "architecture and design". Indeed mixing design with analysis is a sure recipé for project failure. Knowledgeable developers have understood that for decades.

Even those who habitually confuse analysis and design admit that mixing the two makes it hard to evaluate packaged product solutions. Adding the agile practice of incremental delivery of working software, of course, makes buying a packaged product solution impossible. That may be acceptable to a software vendor developing a new product, but it's an instant disqualifier for a user organization that needs a new application system.

Author's response, April 14

The article is not intended to make any statement about whether agile teams favor agile methodologies. Certainly "agile teams favor agile methodology" would be a likely conclusion, but I did not do any research which enables me to make any statement in this regard. Instead the article was intended to communicate that the survey results indicate that 12 agile principles authored more that 10 years ago are still embraced by agile teams. As a researcher, i had no idea how the survey would come out regarding this particular issue. If it had turned out the principles were now outdated and obsolete, that result would have been completely independent of whether agile teams favored agile methodology.

Last modified April 13, 2012

Return to IDI home page
Technical articles