An excerpt from one of Valerie’s answers in the interview about the Linux civility discussions:
I’m one of hundreds of Linux kernel developers, past and present, who agree with Sarah Sharp’s request [for more civility in Linux kernel development] — she’s just the person brave enough to directly call for change from Linus Torvalds and other community leadership. I was a little horrified to see how many top-notch kernel developers spoke up to say that this is one reason why they dropped out of kernel development. So I’m thrilled to hear this will be a topic of discussion at the next Linux Kernel Summit. I hope that other kernel developers will join her in standing up for a working environment without abuse.
I think Linus [insisted on the value of hostile discussion] based on the information he has. For example, he’s probably not aware of research showing that people’s intuition that performance improves after severely criticizing someone is wrong: any improvement in performance is due to random chance, what many people are familiar with as “regression to the mean.” It turns out that when you evaluate the effect of criticism vs. praise on performance scientifically, praise is the clear winner. We as computer programmers should use the same scientific logical approach to community management as we do for software development.
Read more at the USENIX web site.