He visto este clip de Jonathan Blow, y aunque no es el tema del vídeo la conclusión si aplica a cualquier desarrollo.
El vídeo trata de una respuesta a porque él decide no meter test automatizados (o test unitarios). Digamos que para lanzar exitosamente un juego debemos decidir que hacer o que
no hacer, no todo lo que hagamos va libre de carga, hay cosas que nos lastran, ralentizan todo el desarrollo, desde el momento que lo metemos hasta el final. Cuando hacemos un juego hay que tomar 10.000 decisiones, cada decisión es importante, no siempre más es mejor. Hay que actuar con cabeza, porque al final de lo que se trata es de terminar y lanzar el maldito juego.
CitarSo, making a game is all about picking your battles.
It's all about deciding where you put your effort and where you don't put your effort, right?
And so, it's a very deliberate choice that I'm not putting my effort into optimizing the time that tests run because:
* A, it's not really a problem.
* B, let's say you make it faster - that's not free.
It's not free because you've raised the complexity of the situation that you're dealing with every day. Because now there's an extra mode that tests run in, and that has to be maintained for the lifetime of the game.
People who run tests need to understand it. It's one more thing, right?
And games drown you in "one more things."
If you don't watch it, the "one more things" will fucking kill you.
And if you want to learn how to make video games, I cannot overstate the importance of that statement.
It is what you are constantly fighting with.
Constantly, constantly, constantly.