One Week Iterations

Steve Eichert explains how his team does one-week iterations. It sounds like an interesting, fast-paced, and dynamic approach to agile software engineering. There seem to be a couple of unknowns though in this schedule. For example, he mentions they only reserve a couple of hours on Friday to fix any open bugs.

Is that always enough time?

What about bugs you don’t catch and show up in a later iteration? I guess a subsequent iteration could always be dedicated to just fixing bugs and if you are doing TDD then you bug count should be low. Also, is it always possible to have demo-able software in a single week?

For example, what is done if you spend several weeks building some framework libraries? What if you are building a highly distributed system that is mostly middleware and the moving parts are mostly state driven workflow type components? If that is the case, I can see where a little bit of the “Big Up Front Design” might make sense, at least for the foundational level of the system and for the first several weeks at least you may not have anything worth demo-ing.

Tags: agile, software engineering, iterations