The metrics debate is well tread ground.
I’m considering a recent post by Pradeep Soundararajan on measuring test coverage.
It’s a clever response, but I don’t think it’s perfect. What is? It made me think about something I’ve been noticing a lot lately. I really, really, like measurable, testable claims.
I’m taking a course through my university on Organizational Behaviour. It’s part of an IT Management program I’ve begun. I frequently find myself chafing at what I’d describe as “soft” theory. Theory built on observations and assumptions, but unprovable. I’ve had to force myself to recognize that it can still have value if it shapes our perceptions and approaches, although it’s dangerous if taken as dogma.
Metrics. I want things to be measurable, testable, and provable wherever possible. At the same time, I’m also incredibly uncomfortable with metrics, and the way they shape testing. What we measure, and how we measure it, can be an extremely political act. This is a topic that, as I said earlier, has been hashed out repeatedly in the industry.
So I suppose there are two things out here:
-I want to avoid untestable assumptions wherever possible.
-At the same time, I’m trying to learn to find the value in things that aren’t provable and testable, because they CAN still have value.
I’m still really ambiguous about where I sit between those two.