Agile2010 retrospective: Scott Ambler’s “Agile Mythbusters” presentation

Agile 2010, Aug 2010 - 07As I mentioned in yesterday’s blog posting, the recent Agile 2010 conference had 227 sessions, with 180 speakers. Even if I had somehow managed to attend all of them (a complete impossibility, since there often six or seven sessions running in parallel), I wouldn’t have the time, patience, or energy to report on them all. However, there were some interesting ones, and the first one I want to bring to your attention is Scott Ambler’s “agile mythbuster” presentation.

Scott is Canadian, but I forgive him for that — just as I hope he’ll forgive me for not being a native of the frozen northern territories. More importantly, he’s the Practice Leader Agile Development at IBM (whatever that means), and he’s a well-respected guru in the agile community, having authored numerous gooks on object-oriented development, agile methods and techniques,, the Unified Modeling Language (UML), and Rational’s Unified Process.

The main theme of Scott’s presentation is that the agile movement is saddled with folklore, myths, rumors, gossip, half-truths, and exaggerations — some of it created by people who have never had an agile thought in their lives, as well as some who witnessed some of the early agile writings and activities nearly a decade ago. So, you’ll hear people loudly proclaiming that, “agile doesn’t work on large projects!” or “agile doesn’t work unless everyone is physically colocated in the same room!” or “agile doesn’t work on projects where there are rigid regulatory restrictions on documentation and approvals!”

Some of these pronouncements were probably never true; others may have been mostly (or even completely) true back in 2001. But it’s been a decade since the Agile Manifesto was signed, and lots of people have had time to evolve their ideas, mature their practices, and refine their initial practices. If nothing else, lots more people — arguably, many more than the 1,400 people who attended Agile2010 — have been attempting to do what a relatively small group was espousing back in 2001.

But that doesn’t really tell us very much about what works, and what doesn’t; it doesn’t give us any ammunition to confronting some of the more insidious rumors and myths… and that’s where Scott’s “myth-buster” concept comes in. It’s actually quite simple: in conjunction with Dr. Dobbs’ Journal, he’s been conducting a number of surveys during the past few years, to ask programmers, software engineers, project managers, and “agilistas” what their actual experience has been with regard to various agile issues. You can find one useful list of links to about 25 such surveys here, and you can find the results of a fascinating November 2009 “agile at scale” survey here.

Scott is intellectually honest enough to acknowledge that all of these surveys have “challenges,” otherwise known as weaknesses and limitations. The survey participants are generally self-selected individuals with strong feels about the questions being asked, there was little or no opportunity to validate the accuracy of their responses, and the number of people responding to the survey was typically in the low hundreds. Scott summarizes these challenges here on his website, and it’s mostly a reminder that you have to take all of the survey results with a grain of salt. That being said, I find that the results are useful, and they provide a good overview of the general trends and opinions in the agile world.

I’m certainly not going to regurgitate all of the survey results that Scott presented at Agile2010. You may be able to track down a copy of his presentation somewhere on the conference website or Slideshare; but the most direct source of information is the survey data itself, at the websites mentioned above. Meanwhile, I’ll list just a few of the summary tidbits that Scott mentioned, as a “teaser” to persuade you to get the complete set of survey information:

  • 69% of organizations said they were using agile in a Mar 2008 survey, and the percentage had risen to 76% by Jul 2009.
  • By 2009, 77% of agile projects reported they had teams with less than 10 people … but that means that 23% had more than 10 people.
  • In 2009, 13% of respondents said they had succeeded with agile projects that involved Sarbanes-Oxley, while 9% succeeded with HIPPA, and 2% with FDA.
  • In 2008 and 2009, approximately 10% of respondents said they had succeeded with a combination of agile and CMMI.
  • 42% of agile teams are colocated in the same space; 17% are in the same building; 13% are within driving distance; and 29% are “very distant” from other team members.
  • Contrary to popular myths, 90% of agile teams report that they do do some kind of initial modeling and some kind of up-front architectural design.
  • 56% of agile teams do write documentation, and 37% report that they do not.
curriculum vitae.