December 27, 2003
Darwinism and the GNOME Project
Observing and thinking about the somewhat chilly reception Ali Akcaagac's concerns have received on the GNOME foundation list got me thinking about the Darwinism inherent in a big free software project.
It has been said, I believe with some accuracy, that free software is a meritocracy based one's ability to write code (or do other work - Lord knows the actual code I've written has no merit). But looking at the patterns of the evolution of GNOME as a group of individuals, and thinking about Ali's problems, leaves me wondering if there is also some sort of selection pressure that favors people who are nice.
There's a whole scientific literature on the evolution of cooperation, and how cooperative behaviors can be seen to endure in populations when there is such a strong selective pressure for the individual who defects and eats all the group's food. Cooperation obviously did evolve as a human behavior, but it's poised on a narrow ridgeline with the abyss of defection always nearby. (Oh, gawd, did I write "abyss of defection"? Somebody confiscate that guy's thesaurus!)
But in a project like GNOME there is very little defection benefit, so the selection pressure would naturally favor people who are cheerful, cooperative and good at working together. Is it therefore inevitable that people who are gruff and abrasive would get nowhere in the heirarchy, and a community of nice people would emerge?
Posted by John Fleck at December 27, 2003 09:20 AM
It is possible. But it seems more likely that the selection pressure would favour those persons that are best at writing code (or whatever), or have the greatest amount of spare time - however grumpy they are.
If you ask me, the phrase "and a community of nice people would emerge?" seems dangerously close to self-patting. :-)
For instance, I don't think Ali would arrive at the same conclusion.
Ole, I think John is right (and I've thought the same for some time, so perhaps I'm just trying to rationalise my own beliefs)... Mostly because you see exactly the same things in other projects. Consider the relationship between Linus and Andre Hedricks, Tridge and Luke Leighton, etc.
People who are socially... klutzy... tend not to 'make it' in Free Software, because what we do is as social as it is technical. "Nice" is arguably overdoing it (though I think it's accurate in GNOME's case - we almost have a family here), but people who are cooperative, grateful, thoughtful and sharing are generally given more trust, responsibility and acceptance.
If no one trusts or wants to deal with a particular contributor, it's unlikely that they'll get much traction on technical issues, regardless of their coding ability. :-)
I think the selfish gene hyposthesis for altruism is a more apt description of the mechanism at work in OSS. People contribute to the GNOME project because they expect to benefit from the project or software. The leaders are those individuals that provide the greatest benefit to the contributors. All of us can contribute code, translations, documents, and arts, but leadership is only conferred to those that help others make those contributions. I'd argue the leaders are the individuals that contributors trust to make fair and good decisions.
Until I read this I didn't really think about the type of people who are apt to be successful in the GNOME project. However I do agree the nicer and more tolerant people can be more successful. I'd like say Ali isn't very tolerant of the decisions or diversity of people in the GNOME project. He takes a lot too personally when the anger or frustration is more directed with rehashing past decisions and not at the person. However the more tolerant people like Nat, Havoc, and Glynn are the leaders of GNOME. So, I'd like to substitute "nice" for "tolerant" in your analysis.
Jeff, I can see your point - it can only help you if you're behaving nice and friendly. Especially if you want to help coordinate a large project.
I think what I am objecting to is the idea of our community and (free software communities in general) being so special. Of course, since it is basically a gift community where people help each other, everyone tends to be friendly and, well, helpful. But we are still all normal human beings. And in my experience most people are friendly and nice if you treat them with respect, anyway.
It is perhaps correct that there is a selection pressure for the perceived leaders. Like with Linus. But I don't think the same is really the most important factor for everyone else. Those who do the work, get to decide, cf. my remark about spare time.
What prompted my initial comment was the timing of the observations. I don't see the treatment of Ali as an example of nicety - it looks more like group bullying to me.
Not that I'm trying to excuse Ali. I wasted quite some time on him myself when he started the large flamewar last time. He was truly a pain in the back. But I don't think is an excuse for not treating him with some respect. Even if you show people to the door, you can do it with respect. :-)
You're right, "nice" is definitely not a neutral enough word for this, but while value-laden, it's a helpful shorthand for the more complex set of cooperative behaviors that are rewarded in our Darwinian culture. It's not that GNOMErs are better human beings, it's merely that a certain set of behaviors tend to be rewarded in this environment, and so persist, while behaviors counter to that get stomped upon. I think your example of Ali's past behavior precisely demonstrates my point. Many people have apparently had that same experience, that he was for them "a pain in the back". That is the dynamic I'm talking about here, a set of behaviors that are not rewarded in our little Darwinian petrie dish.