As a Canadian, I can be polite to the point of pathology. Yet my reaction to the discussion of politeness on the Linux kernel list is decidedly mixed.
Partly, my own sense of politeness makes me think that anyone not involved in kernel development has no business having an opinion on the topic-- which means that, by my own standards, I am being rude here. More importantly, everyone involved is flinging half-truths the way that monkeys fling feces, and seem to be ignoring the core of the matter.
The topic has flared up again because Sarah Sharp announced that "I am no longer a part of the Linux kernel community," and is winding down her involvement. Her decision was apparently months in the making, but she considers kernel development a "toxic" atmosphere, tainted by rudeness rationalized as "extreme radical honesty," sexism, and homophobia. "I feel powerless in a community that had a “Code of Conflict” without a specific list of behaviors to avoid and a community with no teeth to enforce it," she writes -- and who can blame her?
Sharp's blog was followed hours later by Matthew Garrett's announcement that he was forking the kernel. He notes, probably correctly, that, as "terrible and touching" as Sharp's comments are, they are unlikely to change anything. Then he adds that he is forking because Linus Torvalds rejected his idea for the future development of the kernel that he wants to pursue.
It is hard for even non-Canadians to argue against politeness. Politeness is a universal value, and in most circles the only possible response to comments like Sharp's or Garrett's is uncritical agreement. Nor can the description of kernel development as a toxic environment be seriously questioned by anyone who has followed it. Off-hand, I can think of at least four kernel developers who have resigned over incivility, so the issue obviously needs to be addressed.
All the same, after months of campaigning, Sharp apparently never noticed that she comes across as an upstart, delivering unwanted lectures -- that she can, in fact, be rude herself. Similarly, is Garrett taking a principled stand, or hijacking the issue of civility to camouflage his annoyance at not getting his own way? At times, their egocentricity seems second only to Linus Torvalds', who persists in responding personally rather than as the leader that people have every right to expect him to be.
Just as importantly, I suspect that the picture of Torvalds that Sharp and Garrett are promoting is more of a caricature than a photograph. Granted, Torvalds can be as rude as they say he is, yet logically their picture must be incomplete. The impression he makes face to face or on stage is so self- effacing and mild that it is hard to reconcile with the tyrant of the mailing lists, but both appear a part of him.
Moreover, nobody becomes a leader of a major project for so many years by consistently being unpleasant. If they were, they would be deposed in a matter of months.
Perhaps Torvalds merely has the sense to hire lieutenants with the people skills he lacks? Yet with the exception of Greg Kroah-Hartman, I have never heard that his lieutenants were outstandingly agreeable. For all his flaws, there seems more to Torvalds than the current controversies would allow.
None of these comments are meant to suggest that Sharp or Garrett are wrong, or to excuse Torvalds. However, what surprises me most about the issue is that no one has said the obvious: namely, that what is happening is a clash of women's culture and a particular sub-culture of men.
The male sub-culture that I am referring to is mostly unknown to women, except as its victims. If you are a man under about forty-five, or your work experience is limited to academia or white collar positions, you may have never encountered it, either.
In this sub-culture, men are continually sparring verbally. They insult each other constantly, and never miss an opportunity to denigrate each other's ability, intelligence, or sex life. If you are an outsider, you might expect violence at any moment. Yet the men in this sub-culture can sometimes be mutually supportive, and, in an actual emergency, the men who are the most abusive can be the most generous.
If you are uncompetitive, this atmosphere can be devastating. However, the key to surviving is the realization that those who abuse you expect you to answer them the same way. Shy away from the verbal attacks, and you risk becoming a target of bullying.
This sub-culture is not the only one available to men. If anything, it has become less common in the last few decades. But what matters is that it is the culture that the Linux kernel project, with its majority of men, has unconsciously adopted. Women or younger men can adjust to the culture, often equally unconsciously, but doing so can become a strain, conflicting with their conscious ideas about proper behavior.
The outsider is soon snared in a double-bind: participate and come into conflict with your own principles, or refuse to participate and feel like a perpetual victim and an outsider. Either way, they can never win.
With more women entering free software, the disappearance of this sub-culture is as necessary as it is inevitable. When women contributors become numerous enough -- perhaps as few as one-third of the project --it will disappear as inevitably as the pornographic screen savers from an integrated office.
Meanwhile, to focus on the abuse is to mistake the symptoms for the problem.