A view I share with a small but growing number of people is the idea that software releases are experiments. An experiment needs a hypothesis, and that hypothesis needs to be falsifiable – ot…