Testing The Right Things

Dec 9, 2013

I’ve been thinking about the next level of TDD for a while. The current (most common) model of TDD doesn’t really work as well as it used to. Rather than thinking about how the application should function, I prefer to think about how the application should function and how those functions should be built.

I prefer to sit and think about the architecture. I ask myself questions that actually drive out the design. I think about what objects I will use and what messages they will send. Are the methods private or public? Does this data need to persist? What are my edge cases and how will I handle them? All of these questions guide my tests. These are my test cases.

The Just Test Everything Approach

Today’s model of TDD promotes just diving in once you know how something should behave. Start everything with a failing test. Literally, everything. Even loading the index page.

It’s easy to test for everything, but is it a waste of time?

TDD is awesome, but I feel it drives many away from just sitting and thinking. Rich Hickey is really onto something with Hammock Driven Development

The Key is Unit Testing

Lately, I have been trying to strictly unit test everything. If I am using a framework, testing routing or page loads seems inefficient. It’s already been tested by the framework and by people smarter than me. I try to move everything I am testing to logic that I introduce. This really leads to a more modular application with lots if classes that I have defined. My controllers are cleaner and so are my tests.

By building classes for everything, I am have the opportunity to really think about how the parts interact with the system.

Listen To Your Tests, But Think About Them First

Don’t let TDD tell you what code you should writing, but instead, you should already know what code to write; TDD should drive functionality. I like to know what I am building before I am building it. The model of TDD that newbs follow promotes having failing tests tell you what to do. Maybe this is the wrong thing to tell them. I think it’s better to have newbies think about the solution they have come up with and test that.

Think about how dangerous it is for someone new to programming or a language to start programming something without knowing what they are programming.

Think, design, and then test.

If you enjoyed this and would like to be notified the next time I write something, you can subscribe with your choice of RSS or Email.

I promise not to spam you and you can unsubscribe at anytime.

Also, here are some other items you might like longer longer:

or go see everything in the archive.