Off the Beat: Bruce Byfield's Blog

Every few months, someone accuses Linus Torvalds of being abusive and rude in his leadership of the Linux kernel. Yesterday, it was kernel developer Sarah Sharp's turn, in an attack that was as noteworthy for its idealism as its deplorably poor tactics.

The discussion began on the kernel mailing list when Torvalds suggested, perhaps facetiously, that lead developer Greg Kroah-Hartman might benefit from being less polite.

Sharp intervened with a critique of Torvald's own behavior, telling him that "you're one of the worst offenders when it comes to verbally abusing people and publicly tearing their emotions apart. " 

In response, Torvalds defended his tone as a tool for getting to the point, claiming that he reserved his harshest language for veteran kernel developers who did stupid things when they should know better. He  could be polite when necessary, he insisted. However, he did agree to participate in the discussion suggested by Sharp at the next kernel summit.

Coming from Canada, a country often ridiculed for its politeness, I naturally sympathize with Sharp's position. Contrary to Torvalds' insistence, honesty and effectiveness on the one hand and politeness on the other are by no means mutually exclusive. In fact, if you want people to learn and not waste time being angry at you, politeness is generally more effective.

Moreover, Torvalds' aggressiveness is a very masculine style. Not that men like facing it any more than women do, but they are more likely to have experienced it, and to have learned that the way to cope with it is to respond the same way.

To give him credit, Torvalds seems to expect answers in kind, and not to hold grudges when people stand up for themselves, as shown by his humorous response to Sharp's first salvo: "That's the spirit. Greg has taught you well.  You have controlled your fear. Now, release your anger. Only your hatred can destroy me. Come to the dark side, Sarah. We have cookies."

However, although Sharp has obviously learned how to respond, other women may only be intimidated. For this reason, it seems at least a working theory that Torvalds' style discourages women from entering kernel development. As lead kernel developer, Torvalds obviously influences the tone of the entire community, and a change in his behavior would soon become the norm among kernel developers.

Tactical errors
At the same time, a mischievous side of me can't help noticing that Sharp, who is as much known for her feminism as her kernel work, is making the same tone argument that is often used to discredit feminists.

Just as feminists are told that they would receive a more sympathetic hearing if they were more pleasant, so Sharp is implying that Torvalds would manage the kernel better if he exchanged his bluntness for politeness. Not only does this seem a double standard, but in light of the Linux kernel's outstanding success, why should he listen to the pleas for politeness?

Moreover, as I read through the discussion, the rudest person appears to be Sharp. "Not *fucking* cool," Sharp says at one point. "Yeah, just try yelling at me about this. I'll roar right back, louder, for all the people who lose their voice when they get yelled at by top maintainers"

Meanwhile, while being accused of being impolite, Torvalds is explaining himself at length -- something he is under no obligation to do. "And yes, I'll happily be part of the discussion at the [kernel summit], he adds. "But I think you also need to be aware that your 'high horse' isn't necessarily all that high." Although Sharp does joke at one point in the discussion, on the whole she comes across as far more confrontational than Torvalds.

However, an even more basic question about Sharp's tactics remains: in what sane universe does a relatively new contributor publicly berate a project leader and expect anyone to listen? The fact that she can do so without consequences speaks favorably of the kernel community, suggesting that its dynamics are far more complex than Sharp claims.

In fact, when Sharp proclaims, "I'm not going to put up with that shit any more," observers might legitimately wonder why she thinks anyone should greatly care, let alone what she can do.

But perhaps the biggest trouble is that Sharp's choice of venue means that she is setting her cause up for defeat. Participants on the kernel list are used to listening to Torvalds about general policy, and not in the least to Sharp or anyone of comparable experience. True, one or two people express agreement with her, but since their influence is no greater than hers, their support counts for little.

In the end, Sharp herself suggests the discussion needs to be conducted face to face,. However, the truth is that she might have been better off not raising the issues in the first place, instead of making her point appear defeated or stalled almost as soon as it was made.

No doubt Sharp was angry, and raised the topic without thinking of the consequences. All the same, if she really wanted to put the topic on the kernel community's agenda, she would have been better to stay silent until the summit. In the meantime, she might have lobbied like-minded kernel developers and perhaps come up with a draft code of conduct, thereby making the topic a community issue rather than a personal one.

Now, I'm second-guessing, of course. But the point is, if the issue matters -- and I think it does -- then how it is presented and the timing matters, too.

Belatedly, Sharp is seeking support on social media, but I suspect that she may already be too late. Already, I suspect, poor timing has reduced the discussion to the latest in the long line of criticisms of Torvalds' behavior, condemning its points to be forgotten until the next critique of the same kind is made.