Chaotic Agility

Three quarters of all software projects fail. Why? Because we're still trying to manufacture software instead of create it. Software development is more art than science. As such, it can't be predicted accurately and is difficult to control. Yet how do managers typically respond to a failing software project? More control.

Three quarters of all software projects fail. Why? Because we're still trying to manufacture software instead of create it. Software development is more art than science. As such, it can't be predicted accurately and is difficult to control. Yet how do managers typically respond to a failing software project? More control.

Complexity science has a surprising amount of relevance to agile software development. Developers tend to behave like agents in a Complex Adaptive System (CAS). These agents must be allowed to self-organize to the edge of chaos, which is where remarkable things happen. Most of us agree that agile works. But why does it work? Examining the science behind agility can tell us why.

Once we understand why agility works best for software projects, we will stop trying to predict what can't be predicted and control what can't be controlled. We will allow software to emerge on its own instead of trying to manufacture and control it.

About Matthew Bass

Matthew Bass is an independent software developer, entrepreneur, speaker, and writer. He has over ten years of experience across a diverse set of technologies and has worked at places like SAS Institute, the world's largest privately held software company. An agilist from the very beginning, he continues evangelizing and experimenting with pair programming, test-first and behavior-driven development, and continuous integration. Matthew has spoken at several regional and national software conferences and regularly writes for publications like InfoQ.

More About Matthew »