How to Get Started in Open Source (as a Developer)

This was an email I wrote a few years ago in response to someone asking how they can get started in open source, if they're already a professional software developer. Someone else just asked a similar question, so rather than copy-paste from my sent messages, I'm copy-pasting onto my blog.


If you already work as a software developer, there's a very good chance you're interacting with open source dependencies on a somewhat regular basis. If you want to get involved with contributing to open source, the simple (but somewhat unhelpful) answer is: just start doing it.

To explain somewhat what that actually means, there are four main activities in open source:

If you already a professional software developer, then all four of these are things you can contribute to as part of your normal work, and ideally it doesn't need to require extra labor (or at least, not more than you benefit from). Your job is to create and improve the software your employer uses and/or sells, so it's only a slight mental frame shift to see how "contributing to open source" is a part of that work.

For the first, creating things, practice noticing the thought in the back of your head while working in software, "there should be a module that just solves this problem for me so it's easier". Sketch out how you'd use it, think about what it would do. Then write that module and publish it, and use it to solve the problem you had.

For the second, fixing things, there's a lot written on how to improve your debugging and issue reporting skills, but the most important thing is building the skill of creating small reproducible test cases (ideally, even by adding a test to a program's existing test suite that fails due to the bug). As a maintainer, good reproducible bug reports are always a gift.

Many projects have an ever-increasing list of bugs that have yet to be addressed, so you can always start in on those. But the easier approach I've found is to build a practice of noticing when you think "fuck this, this thing is broken, I'll use xyz instead", and before you do switch to the other option, make a small program to prove it's broken in the way you think.  Then fix the broken thing (or at least, report it to the author with the small proof you wrote). Surprisingly often, you'll find that your minimal test program doesn't break in the way you observed in your actual program, and you learn something interesting about your own assumptions.

The third, improving existing things with new functionality, is a blend of the first two. Keep an eye out for cases where you're using something and find that it's lacking some functionality that seems to "fit" with what it does, maybe even where you're surprised it doesn't do what you need. Sketch how you'd use the added functionality, what it would do, etc. Then add that functionality, or at least request it as clearly as possible, with your reasoning and a clear explanation of your use case. (The author will often have feedback that can be very helpful, such as workarounds, reasons why it's not a good idea, better approaches, etc., so it's a good idea to at least request the feature and see what they say, especially if digging into the code is challenging.)

For the fourth, helping users, there's a lot of ways that people mean to help, but some of these are sometimes unhelpful. When reporting issues you have, you'll be right there in the bug tracker, and will probably see many open issues that could benefit from a minimal reproduction case, or where a requested feature is not fully articulated. Get in the habit of asking questions and clarifying the limits of what you actually know, rather than stating authoritative answers, especially if it's not your project.

There's certainly no moral or ethical requirement that you as a worker personally contribute your time and attention to open source just because you use it in your day job. If it's fun and rewarding for you, if it fills your heart with joy, or if you see some personal benefits from learning or improving your skills, then yeah, just do it. If not, don't. But always know that it's optional, and available to you.

Even if you don't do these kinds of contributions (which is 100% fine!), it is important to encourage your employer to dedicate time and money towards the open source projects that their business depends on. Many companies want to contribute back to the communities they benefit from (for self-serving as well as altruistic reasons), but often managers and executives lack an understanding of specifically which projects matter to the business, and thus where their contributions should be deployed. So it's on us to act as technical advisors.

And if your employer has a problem with you improving your skills by participating in the oss communities that benefit them, and improving the state of the craft that their business depends on, well... find another employer, because that's really a misguided and short-sighted point of view.