I just got back late last night from Surge 2014 and illumos Day, which immediately followed Surge the next day. There were some great talks going on, which I’m sure I’ll also be writing about. But the first speaker in particular dropped something on me that’s bothering me, and it should bother pretty much anyone that hears it.
Garrett D’Amore, founder of the illumos project, crawled through all of the commits and made a really interesting discovery. This is a four year old project, and remains relatively obscure (though some very visible things have come out of it, like zfs). In those four years, about 150 unique contributors have committed code into illumos-gate, the shared core of the illumos ecosystem that distributions are built on. Now on the surface, this number sounds pretty wicked cool. illumos is a fairly unknown project, sadly, so to score commits from 150 engineers sounds like a really good thing. Or is it?
Of those 150 unique commiters, 0 of them were women.
Zero. Zilch. Nada. None.
While we all know that, for whatever reason, software development has developed a sad stereotype of being almost exclusively a male pursuit, we’ve made some strides over the last n years in being more inclusive on the gender spectrum. We’re seeing more women not just participating in technology, but leading the way. We get a lot of value out of having women involved in software, not just in the doubling of the size of the potential talent pool, but in having greater diversity of thought and perspective in how we solve tough problems together.
While we’re definitely still far from seeing a 50:50 blending of male and female engineers on software projects and at technical conferences, there is some movement and it’s good, and we need to keep getting better at it. But the dirty little shame of illumos, now brought to the surface by Mr. D’Amore, is we’ve managed to attract none of them into committing code to the crown jewels of the illumos ecosystem.
Now, to the credit of the community, the ~20 people in the room at the time immediately stopped the forward momentum of progressing through the slides and engaged in passionate dialog about contributing factors, a little bit of grasping about what we could do about it. We don’t really understand the full scope of the problem yet, so it’s hard to really identify effective solutions.
Bryan Cantrill of Joyent spoke very passionately about assholes being the bane of the Open Source community, and how effective they are at chasing good people away from the community. Part of the solution, it would seem, is to not allow alienating social behavior to enjoy influence. This line of conversation grew legs, and ran around the room a good bit. We’ve got some more to talk about here.
There was also a good deal of agreement that the illumos community is a friendlier place for new contributors to come and find ways to get involved than many larger (and more prominent) projects. As someone who has sort of nipped around the edges of this community, I have to agree, I’ve found it really easy to get help in a very collaborative way. What we’ve not done a good job at is marketing, at sharing this with people outside of the community, which is part of why I feel it’s important to tweet about what we’re doing, blog about it, speak, etc.
There was a sense in the room that we couldn’t really quantify, but that several openly suspected was there, that women have maybe tried getting involved in the Linux community and several (or maybe many) have been alienated by some of the bad behavior that can often go on in those communities. The question posed, but not yet answered, was how do we get them to come hang out with us and see that we want them to feel welcome and valued here?
I don’t think we really have the answers at all yet. I don’t think we have even asked the right questions yet. But this conversation is crucial to the survival of illumos, so it needs to continue, and it must result in real improvements to our gender diversity.