You and Jimi Hendrix

This post originally appeared on the Software Carpentry website.

I had a discussion a couple of weeks ago about software development tools and processes with some undergraduate students I'm mentoring. They asked why I'm so finicky about putting things under version control, writing unit tests, and creating tickets to keep track of what still needs to be done. The short answer is, because that's what Jimi Hendrix would have done.

The long answer goes something like this: if you want to be able to improvise without falling flat on your face, you need to have rock-solid technique. Hendrix couldn't have done what he did with wah-wah pedals and overdriven amplifiers if he couldn't first play scales and arpeggios; as one "artist" after another has shown over the last forty years, people who try to do the former without mastering the latter are just making noise. The same is true of jazz greats like Parker and Coltrane, and of classical musicians like Yehudi Menuhin.

I think it's true of programmers as well. I don't think I code nearly as well as Hendrix played guitar, but I know people who do. They don't actually put everything they care about under version control, and they certainly don't always write unit tests before writing code. However, they're fluent enough with those practices to decide when not using them is the right choice.

More importantly, good programmers have done things the right way for so long that they revert to good practice out of habit when they're stressed and tired. That's the real reason I push novices so hard to do things the right way every time: when the deadline is just hours away and nothing is working, throwing away your workflow and reverting to feral coding is exactly the wrong strategy. Practicing arpeggios might not seem like fun when you're starting out, but getting the fundamentals right will make anyone a better programmer, just as learning how to play a twelve-bar blues will let pretty much anyone sit in with the local bar band.