Blog

Two problems with context-driven testing (Markus Gärtner)

On July 3, 2012, in Syndicated, by Association for Software Testing
0

Over the course of the Let’s Test conference in Runö, Sweden, I noticed a problem with context-driven testing. In the past one or two months this turned into two problems I see with context-driven testing. I finally decided to put them out there for further discussion. I hope a lot of you don’t agree with me – and I hope a lot of you folks speak up.

Systemic effects

The most pressing issue I have starts with the systemic effect. I will phrase this as arrogantly as I can. The term “context-driven testing” – for me – includes the notion that we are driven by the context, and the context of our situation is the only way we can think of as we should be handling now. We are testing alongside our context, right? What could be wrong about that?

This interpretation of “context-driven testing” ignores pretty much a central point about context:

The context includes the tester.

Our behavior leads to changes in the system that we call “context”. This is a variation of the Chinese proverb: “When you point your finger to another person, look where the other three fingers are pointing to.” By acting based on “the context”, we influence “the context”, we change “the context”, thereby we will drive “the context” to another status, to another “context” that will change our behavior. To put it clearly: Stating “I as a tester am only driven by the context of the people around me” is foolish, short-sighted, and ignores the systemic effect that our own actions have to the situation, the context of our project.

If we do a poor job of testing, this will influence the stakeholders and project managers around us to lead to different actions upon us. If we do a decent job testing for the stakeholders’ expected results, this context might drive us into different follow-up actions. But rejecting the premise that we as testers influence that context, and are only driven by it, is rejecting the systemic effect that we bring to the party. We neglect the effects that we make upon the project, thereby we will miss a large part of “the context” that we claim that drives us. Ouch!

But there is more to it.

Victim behavior

Victim behavior? What’s that? Unfortunately we have to dive into transactional analysis for this. According to the theory behind the Karpman drama triangle we play a game here. A game with roles as the persecutor, the victim and the rescuer. Huh? What’s that?

People play games at work and while interacting with each other. The Karpman drama triangle puts three roles into context, and I see them a lot. The persecutor is the one evil figure. It’s the devil in the Holy Book, it’s Al Kaida in Western News Reporting, it’s the programmer and the project manager in modern programming (I made this third one up).

As a tester, now, we are victims of this situation. We are helpless there just as the US is helpless against Al Kaida, the German soccer team is helpless against the Spains or Balotelli, and the only hope against the devil is the belief in God. We testers are the victims of the situation, the victims of the context, that became our persecutor. I hope by now you notice how this is an extension to the first point that I made with the ignorance about the systemic effect.

Now, in the triangle, there is someone who rescues us victim testers that are persecuted by the evil context.By playing the role of the victim, we are seeking for the rescuer, whoever that might be. I won’t hypothesize on this; you’re a thinking tester; I think you can make up your mind about who your rescuer could be on your own.

Now, here is my call: Stop being a tester that claims to be “driven by context” as a false victim behavior. You are influencing the same “context” by your behavior, and you might find yourself seeking a rescuer that will not hear you.

I agree with most of the context-driven literature, most of the context-driven teachings, heck, I even claim to teach context-driven testing to others. And I think the name is at least misleading in these two particular points that I described. We are not victims as testers, and we are active actors within the context that we claim that influences us. Deal with it. Now.

PrintDiggStumbleUpondel.icio.usFacebookTwitterLinkedInGoogle Bookmarks

 

Comments are closed.


Looking for something?

Use the form below to search the site:


Still not finding what you're looking for? Drop a comment on a post or contact us so we can take care of it!