There's a highly interesting post by Ben Frederickson, titled Why GitHub won't help you with hiring.
It makes some very good arguments against requiring a strong GitHub profile for job applicant screening. The main argument is that most great developers don't have one. I think the best argument against it is that unless you hire people to work on open source code, requiring them to have a track record of contributions to open source software is hypocritical.
However, it's not my point. My point is that his analysis is hinting at the true size of the open source developer community, and it's not as large as it seems.
From the GitHub archive data, he got these figures:
- Only 17% of GitHub users pushed any code in the last year
- Only 7.4% of GitHub users pushed more than 10 times
- Only 1.4% of GitHub users pushed more than 100 times
- Only 0.15% of GitHub users pushed more than 500 times
As of 2020, GitHub is the largest code hosting platform with about 40M users.
Now, let's convert Ben's relative numbers to absolute values:
- Only 560 000 people pushed more than 100 times.
- Only 2 960 000 people pushed more than 10 times.
So, if you somehow bring them all together, the community of people who ever made some code dumps or drive-by contributions is about as large as the populations of Amsterdam, Athens, or Chicago.
The people who actively maintain projects could fill a city like Dresden or Las Vegas.
That's compared to 40M registered accounts. It's hard to tell how many of those accounts are duplicates, or how many programmers are in the world.
Now, let's make an incredibly conservative estimate: most accounts are duplicates and there are 10M programmers in the world. It means only one in three ever helped an open source projects, and one in five maintains a project of some kind.
And that's when almost every programmer depends on at least some open source code for their daily work.
I think this leaves no open questions about whether companies should allow employees work on open source in their paid time, or whether we should advocate for participation. The answers to all those questions is "yes".
Top comments (0)