As a technical agile coach and trainer, I help teams discover ways of testing. Some teams ignore tests altogether, while others write every possible test possible, wasting valuable time and not being able to deliver at a good pace.
My first question is always this: How much does the customer pay for tests?
That’s right! Not a dime. I don’t even ship my product to them with any tests. They aren’t even compiled into bytecode for them. They are not going to pick up my application and open a debugger to make sure I’ve written tests that pass. They don’t care how many tests I’ve written or my code coverage ratio. They don’t care about unit, integration and acceptance tests, or how much time I spent on mocking and stubbing to isolate my functions. They only pay for working software.
So why write them?
I don’t write tests for the user. I don’t write tests for management. I write tests for me. I write them for my future self. I write them for my team members and any other developer that will need to change my code.
I write tests to prove that what I have written is what I have intended. I write them to make my code manageable, to help me refactor when, inevitably, a new feature or change request arrives. I write tests so that I can fearlessly alter a system and know what I will break, and to find and repair bugs quickly before they are pushed into production.
I write tests so that my team members can feel a sense of code ownership, so that they too can alter, improve and remove my code and be able to predict the outcome. I write them so that it becomes a form of documentation of the capability of the system.
Certainly they take time to write, but they save all kinds of time when it comes to changing things later. They allow me to do the one thing that software needs to do in the rapidly evolving market: adapt. I can adapt quickly to the needs of my customers to deliver quality features rapidly.
I write as many tests to make myself and my team feel confident that we can continuously develop a quality product at a sustainable pace, responding to the changes of the market and the needs of our customers. I write enough tests that I am releasing nearly bug-free code and I write as many of them as needed to satisfy just that!
In my Agile Software Developer training events, I help developers learn ways of writing tests to improve the quality of their work, so that they are able to spend more time developing new features rather than debugging old ones.
[This article was originally published on Agile Advice on 28-Feb-2019]
If you find this useful, please consider contributing with our
“Value for Value” model.