Tuesday, 1 May 2012

How Bad Code Commit Habits Hide Problems

Commit habits, it seems, are highly personal. I know people who commit incredibly fine grained work, and I know people who commit X thousand changed lines regularly. Who am I kidding, I just did that yesterday. And last Friday... There's good and bad in everyone's behaviour, and I'm interested to think about what's what in this particular case. Most importantly, it seems to me (and others), that bad code commit habits can cause terrible de-coupling of the inputs and outputs of the development process from the coding itself.

I was initially inspired to write a few words on these matter by this post, and I've been reminded to finish it by the new tester on my team, Sona, who's come from a class IV medical devices background, where their auditor insisted that every single commit was linked to a defect or change request ID.

Part of me wants to pursue this at Vicon, and part of me knows that I'll get a hard time doing so, so where's the middle ground?