The topic for this “Software Development Mantras” post is “defensibility.”

First, let me introduce the concept of “TIMTOWDI” (pronounced “tim-toady”). It’s an acronym for “there is more than one way to do it,” and software development may be unique in that for any given task, this holds true.

So, given N different ways of implementing the task at hand, which one is better?

If one way is dramatically simpler (either conceptually or implementation-wise), that may be guide your decision process.

If one way is especially more testable, that may indicate better separation of concerns, or at least lead to a provably correct solution.

I’ve found there’s a simple litmus test that is always applicable, though. Just put yourself in the developer’s shoes several months from now, who needs to work on this codebase, and ask yourself: “which solution is the most obvious, least surprising, and most defensible.” It may seem like CYA, but it’s for a good cause — you don’t want the engineers of tomorrow putting your code on thedailywtf!

What other development truisms have you found?