Wednesday, March 3, 2010

There is nothing like starting young

I fully support that test automation plays a key-role for the testing team to cope with rapid changes. But I also see, is that many stakeholders in the software development process have an odd way of understanding the consequence of test automation. It doesn't come to you automatically.

While most management agree that test automation is important, interestingly many have no clue that efficient test automation (and also manual testing) requires the code to be testable.

Testability seems to be a foreign word and it is therefore not understood as being one of the key-roles to successful implementation of test automation. So, why is this important part not pressed ahead?

The opposite is often the case. Making the code testable is considered a low priority "feature" with the result that the effort to do test automation increases dramatically. And then people ask why the test team can no longer cope with the vast amount of changes. I hear phrases like "Oh, yes, that is a nice idea, but we will do this later". Later here means at a time none has the time anymore to do it and when it is too late for the test team.

Some also believe the fairy-tale that test automation saves tester resources.

In all those years being involved in a handful testing and test automation projects I realize an ever recurring pattern in this scene. The number of features grows over time and as a consequence also the number of test cases that should be executed.
However, since the number of testers remains a constant regardless of these facts it ends up with the testers limiting their activities on scratching the surface of the AUT instead of practicing the theories what all those testing certification authorities are trying to teach us.

I am usually sitting in a fast train that never stops and I can only hope that we don't hit the buffer stop with full speed like Toyota experienced recently.

So, where is the link to the cartoon? All evil starts with missing testability, and only today I have learnt once more that I was involved too late to bring in any meaningful testability features which could have helped testing more efficiently.

Hey Linda, when you read this, yes, this is the feedback I provided to your article at StickyMinds only today. By the way, I also read your other articles and you must know that your book was one of the very first I ever read about test automation, many years back.

No comments:

Post a Comment