One size kills all

Yesterday I had a conversation with a senior testmanager about test processes. He was very glad that his (large) company had a defined and structured V-model based test approach. My reaction apparantly wasn't that enthusiastiac, and he asked me why. And I said: "One size kills all". Let me explain that.

When I look at projects, I see a dynamic environment with a lot of uncertainties, change and growing knowledge on what we are developing. One way to deal with that is to introduce a defined test approach. Through this the fuzz becomes more structured and the goal becomes more clear. But sometimes a different approach is more succesful; for instance by allowing the dynamic to exist, to operate closely as a team and to frequently interact with the customer. That way the goal also becomes more clear and the business becomes more aware of what they actually need. For some projects a strcutured test process might be a solution. For other projects that same solution will only makes things worse.

Let us as testers be aware of the fact that one model can never be the solution for all problems. If someone tells you that, then you know (s)he's a liar. Our world, our projects and software development are far too complex for that. And they involve human behaviour too!
I think we should develop mutliple testing models and approaches based on multiple paradigm views, so that we pick a model that fits our situation. Because else we're just fitting the problem to our solution.