The picture above shows Samuel Langhorne Clemens, better known by his pen name of Mark Twain. You may remember him as the author of Huckleberry Finn, and a number of other books. Why is he here in this post about testing on this blog about testing? Well Mark Twain was also the author of this famous quote, which came about when his obituary was published in an American newspaper while he was still very much alive:
“The rumours of my death have been greatly exaggerated”.
Recently there has been a lot in the testing press, online and on social networks regarding the death of test cases. “Test cases are so 1980’s” was one tweet I saw. I looked at this and thought “Why?” Why are test cases suddenly out of date or dead, why is scripted testing suddenly seem as a bad thing? Are people just trying to be cool, thinking only exploratory testing or session based testing is needed and test cases are suddenly obsolete?
I think there is a place for both disciplines and it seems others agree with me. Sure, test cases that add no value, and test cases which are merely endlessly repeated by testers who are thinking more about what to have for dinner tonight than what they are testing, are of course worse than useless. They give you a false vision of quality, a sense that everything is OK when clearly it is not. Where testing should add value and provoke debate and opinion they merely suppress the discipline to ‘button pushing’. But the act of designing a good test case, of spending the time and brain power to ensure that the system under test is exercised in a clear, efficient and repeatable way can only be a good thing. It should not be reduced to something that is perceived to be out of date, stuck in the 80’s or worthless. It’s still a very valuable tool in the testers arsenal, allowing them to get an upfront view of the changing requirements of the system they are testing and to document in a clear and consistent way how they intend to test.
Scripted testing and exploratory techniques should sit hand-in-hand. Both add value. In my experience you need both. An illustration of context driven testing is “The essential value of any test case lies in its ability to provide information (i.e. to reduce uncertainty).” A good test case does exactly that. It provides information. It can help show a system will be testable at a point where that system may not even be is testable. It enables others to learn the system and it provides a way of sharing information between those working on the system and outside of it. No two testers have the same level of experience. Test cases can enable consistency in the areas of a test strategy where that is needed. In the case of regulatory approvals and safety critical systems then they are mandatory anyway.
So let’s not let the test case become an old, out-of-date idea. Let’s make sure that it’s one technique we use as part of a well-rounded test strategy whose aim is to add value to whatever system we find ourselves testing.
