More than XP. More than one day.

Following on from demonstration and exploration and of the idea that TDD and non-TDD codebases seem to have different statistical properties (XPDay 06, Spa 07, Agile 07) , this session explores how to make use of that as a guide to refactoring in practice.

2 parts:

I Recap A quick introduction to the concept that programs exhibit various has scale-free properties, and that the distribution of cyclomatic complexity is one, and that it is demonstrably different in TDD vs non-TDD code.

II Going Further Some attendees at previous sessions on this topic have reported that refactoring has sent the metric the “right way”. An enhancement to the tool for gathering the metric presented previously will allow a drill-down to find hospots of complexity distribution oddness, these will be compared against the well-known code smells and the implications for refactoring explored.

Atendees will use the tool to explore code (their own or downloaded) and discover if the metric can indeed point to candidates for refactoring. They will also look at the metric applied to their code across time and discover if the metric changes in ways consistant with their feeling for design improvement. And if so (and I’m betting that they will), discuss how they can use it to guide their deveopment effort in future.