the test orbit


When approaching testing something there needs to be an up front decision on how proactive and how reactive you want to be. Being fully reactive is ok for a simple product, but as soon as complexity rises you will pretty quickly crash and burn. Being fully proactive is important for a complex product where failure has a high cost, but for a simpler product will mean entering a maintenance hell from which it’s hard to return. I like to think about achieving the right balance, where a small amount of maintenance is required to keep us at an acceptable level of risk of failure. This is the test orbit.

choosing from the infinite pile


The six Venn diagrams of testing. In only one of these do we know the truth about the product (diagram 6). Even in this case, we don’t actually know that we know everything. This illustrates the point that the goal in testing is not to uncover the truth; rather, the goal is to define and follow a strategy (allowing for some alteration as new things are learned). Of course the hope is that these two things align.