Red, Green, WTF?


Haskell, But Quickly

Heads Up!

I'm writing a book about building and deploying web applications with Haskell and Yesod. Want to know when it's released? Click here.

Test-Driven Development is gradually working its way into the front-end development ecosystem, but I don’t feel the adoption rate is as high as it probably should be. There are a couple of reasons for this; or rather, the two reasons I can think of are sides of the same coin.

I’ve been consuming my fair share of TDD-flavoured literature lately, all written and recorded by the revered and outspoken thought-leaders in this space. It’s funny; I could swear they all copy each other’s prose. I’m particularly amused by the phrase “reason about”, as in: “…blah blah blah, I find this approach makes the concept in question easier to reason about”.

The interesting thing here is that in order to reason about any particular concept, the concept needs to be analysed at a reasonable level of abstraction. Most of the TDD discussion tends to fall into one of two levels of abstraction:

  1. An overly-simplified Hello, World! syntax example, or:
  2. Dry academia involving diagrams of boxes with arrows pointing to other boxes

A typical ‘Hello, World!’ example of using a test framework like Jasmine will look something like this:

describe('Hello World', function() {
  it('should return hello world', function() {
    expect(helloWorld()).toEqual('Hello World');
  });
});

What does any of that mean? From the outset, these five lines are confounding. In the first line, we say we are describing a ‘Hello World’. What is a ‘Hello World’? It doesn’t even sound like a noun! In this case, the fact that the example is nonsensical actually makes the concept more difficult to reason about.

Empathise with your fellow developer. They’re thinking: “Well, I have this web form, and some JavaScript that validates the input fields; I wonder how I might test that my JavaScript does what I intend.” In which parallel universe does writing a function that returns the string ‘Hello World’ have any use to our friend writing form validation functionality?

If you’re explaining TDD, starting with an unrealistically simple example steepens the learning curve.

The academic types sometimes go too far the other way. An article on Test Doubles opens with an explanation of stunt doubles and their role in the film industry. What follows is an analogy between stunt doubles and Test Doubles. You could argue for or against that connection, but whilst you’re pondering the difference you’re not thinking about how this applies to your codebase.

I’m convinced the two are equally discouraging to a TDD beginner. Being forced to draw parallels between your codebase and example code with no realistic use-case is an unnecessary added complexity when you’re trying to understand how and why to write unit tests for that code.