9 October 2019
When I talk about the creativity, critical thinking, and exploratory learning that comprises testing, there is a good chance someone will ask "what about regression testing?" One simple answer is that regression testing is just like new testing, except it has a specific risk focus: the risk of bugs introduced to the product due to changes. But that is too simple, because when the question is asked it is not about testing at all-- it's usually about the value of formalized automated output checks. Many people confuse the algorithmic process of output checking with the social and intellectual process of testing.
I have observed the testing industry for 32 years now, and I think this is mostly why testing as a field never earned respect. It's why every testing conference these days is basically an automation conference. (Programmers don't have this problem, because when programmers automate keystrokes it's called "the build process" and not "automated programming.")
Automated regression checking is useful and in many cases absolutely indispensable. While I don't confuse it with testing, good automation is created through good testing and it supports good testing. New testing and regression checking are, in a sense, opposites, but they are in no way enemies. In this talk I will help you sort these things out so that you can approach your work with clarity.