I idolized Torvalds growing up and respect him to this day (I still sometimes randomly say to myself "Hello, this is Linus Torvalds, and I pronounce Linux as Linux!"). I also applaud the strides he's made to improve how he treats kernel contributors.
That said, this is a non-answer and it irks me. The truth is most kernel contributes are likely CHWMs. Instead of this "I can't see color... no literally because of the computers" non-answer he gave, it would mean a lot if he acknowledge that likelihood and outlined what might be done to change it.
I get that convincing people to work on free software is sort of hard to begin with. Also I get that when companies pay full time employees to work on the kernel its usually a boon for the whole community. That said, it would be cool (and free!) if the kernel leadership signaled to those companies that they should take diversity more seriously when they are hiring individuals to work on the kernel.
He believes in meritocracy, that's the answer he gave, or at least the way I interpreted it. Of course he's not going to support what you ask of him since that goes against meritocracy.
Since companies embraced FOSS, most for cutting costs than anything else, there was an increase in POSIX clones with more interesting licenses.
IoT is full of them, NuttX, Zephyr, RTOS, MBed, microEJ, Azure RTOS ThreadX among a few others.
So after the generation that has created and groomed Linux moves on, its golden age will be most likely forgotten, replaced by endless forks and alternative POSIX clones.
'more interesting' is a double edge sword: sure it means that you're not obliged to disclose your proprietary code but other contributors can do the same, so it goes nowhere.
It seems that the GPLv2 is the 'right balance' as shown by the very little adoption of the GPLv3.
Too bad that for libraries we don't really have found such license..
For me this is more of a philosophical question than anything else, but it seems that only in hindsight those pushing for those licenses will realize that shareware and PD are back, just got renamed into community and enterprise editions.
Diversity is good, but when it comes to stuff like the kernel or other system critical stuff, the color of your skin does not make you a better engineer. What do people even mean by diversity anyway? Addition of black or brown people? If a team of all Indian engineers adds a white person to the team, does that make them less diverse? Anyway, we need to be careful about this. Sadly, when management makes a poor hiring decision under pressure to add diversity and things go sour, people end up wrongly blaming the individual.
Torvalds is correct to play it safe here, because signalling for diversity ends up sending a signal to a significant portion of the free software community that the kernel development now prioritizes diversity/race/skincolor over the function and performance of the kernel. And fortunately/unfortunately, the kernel doesn't care what color skin of the hand that writes and debugs its code.
The color of your skin doesn’t make you a better engineer. But a community that excludes a portion of the population does limit the number of great engineers to draw from. Diversity helps improve the community and create a world where people who’d make excellent engineers aren’t pushed away.
That said, this is a non-answer and it irks me. The truth is most kernel contributes are likely CHWMs. Instead of this "I can't see color... no literally because of the computers" non-answer he gave, it would mean a lot if he acknowledge that likelihood and outlined what might be done to change it.
I get that convincing people to work on free software is sort of hard to begin with. Also I get that when companies pay full time employees to work on the kernel its usually a boon for the whole community. That said, it would be cool (and free!) if the kernel leadership signaled to those companies that they should take diversity more seriously when they are hiring individuals to work on the kernel.