Open source developers, it turns out, tend to be well paid. That’s one possible conclusion to be drawn from a recent Linux Foundation report (PDF), which found that over 75% of the top maintainers for the 200 most active open source projects are paid to work on open source full or part-time. This isn’t a new development (I wrote about it back in 2008), but it bears repeating since we are apparently in the midst of an open source sustainability crisis (again).
As Luis Villa has suggested, “getting paid” isn’t the same thing as “comfortable work,” which can lead to burnout. But it does suggest we may need to approach the conversation with more data and less hand waving.
Getting paid to write free stuff
In a world increasingly dependent on open source software, with 80 to 90% of modern applications built using open source, it’s critical that such software be adequately maintained. Starting in 2014, the Linux Foundation created the Core Infrastructure Initiative to help fund and support open source projects deemed critical to global information infrastructure. Such an effort, however, mostly covers for underfunded developers/projects, and isn’t intended to fund open source projects en masse.
What about those?
Well, as part of an initial census into the health of interdependent open source projects, the Linux Foundation dug into GitHub data to determine who’s contributing what to the top-200 most active open source projects. The findings may surprise you:
[A]n interesting pattern [emerges from the GitHub data]: a high correlation between being employed and being a top contributor to one of the FOSS [free and open source] packages identified as most used. Contrary to [a] popular image in open source discussions of “the overworked and underpaid programmer,” an analysis of 2017 GitHub data found that some of the most active FOSS developers contributed to projects under their Microsoft, Google, IBM, or Intel employee email addresses. Even if the contributors to the projects listed in the appendices do not receive direct compensation from private companies to develop these packages, their status as a member of the FOSS community could have endorsed their qualifications for their current paid employment.
My response? I’m surprised that it’s not more than 75% of core maintainers that are employed to contribute open source code. This could be low simply because it’s difficult to correlate GitHub profiles with developers’ real-world personas/employers (the Linux Foundation wasn’t able to establish any sort of company affiliation for an additional 10%, despite efforts to search LinkedIn and other sites to correlate developers with employers). As Linux Foundation executive Chris Aniszczyk pointed out, “Many folks don’t contribute code from a company email address even though they contribute on company time.” Such contributions may directly relate to the developer’s employment, or it may indirectly benefit their employer (or, of course, can simply be a hobby).
At any rate, the finding that open source developers tend to be well-compensated isn’t isolated to the Linux Foundation’s survey. As analyst Lawrence Hecht has noted, a range of independent studies have reached the same conclusion.
And yet…it may be too soon to celebrate.
Though it’s fair to conclude that open source developers, in general, aren’t starving in the streets, the Linux Foundation is in the middle of spinning up a more comprehensive survey of thousands of developers to get more detail. Specifically, “The survey will explore the contributor’s level of engagement, employment history, and employer’s policies on developing FOSS in the workplace.” Such a survey should yield helpful data.
In the meantime, Villa’s point remains valid: “It’s also important to note that it is perfectly possible to have a very well-paid day job most people would love, and to have to squeeze in your work on open source on evenings/weekends/20% time, which is a recipe for burnout even if you’re well paid.” In other words, open source developers may not be starving, but that doesn’t mean we shouldn’t be thinking about “sustainability” more broadly than “earns a paycheck.” Huge pressure is often heaped on maintainers with little compensating empathy (or assistance).
So, is open source “sustainable”? From a cash-for-code perspective, I’m betting the answer is a firm “yes,” and will wait on the Linux Foundation’s more comprehensive survey to confirm or disprove that belief. But I’m equally sure that cash isn’t the only thing that matters here, and we need to do a lot more to make open source development something as emotionally sustainable as it (likely is) financially sustainable.
Disclosure: I work for AWS, but nothing herein directly relates to my employment there.