I’ll reply to myself to avoid editing the above (for transparency).
Linux users in this post: think on all issues that you had with your system. Bugs, papercuts, devs assuming use case, regressions (shit stopping working), dependency hell, anything. How many of those issues apply to the kernel, in a way that you can say “the kernel devs fucked it up”? For me, never.
I have a hypothesis, that I do not know the truth value of, that the kernel not annoying the shit out of us users is directly related to Torvalds’ propensity to tell people “your code is GARBAGE”, instead of sugar-coating it. And that free + open source projects where project leaders don’t do this tend to be crappier. (Does anyone here know a good way to falsify this hypothesis?)
It’s half this, and half an explicit policy “we do not break user space”. Together it meant that if you did anything that screwed up the user space you got told about it at length.
Now Linux culture is established enough that it only really needs the policy, and not the cussing people out to enforce it.
I wasn’t aware of that email, only the quote itself.
…not gonna lie, I think that it was beautiful. I have my bones to pick with pulseaudio but come on, you don’t shift blame like this, the guy deserved some smacking.
On-topic: I have my doubts if policy is enough to enforce it, or at least to enforce it in an efficient way.
Well this is it. What really enforces the policy is rejecting commits that break user space.
Now if you’ve got a large enough group of devs, rejecting commits is fine, but if you’ve only got a small group you need everyone to be working productively, and you can see why Linus ended up giving angry feedback about commits that were wasting everyone’s time.
I’ll reply to myself to avoid editing the above (for transparency).
Linux users in this post: think on all issues that you had with your system. Bugs, papercuts, devs assuming use case, regressions (shit stopping working), dependency hell, anything. How many of those issues apply to the kernel, in a way that you can say “the kernel devs fucked it up”? For me, never.
I have a hypothesis, that I do not know the truth value of, that the kernel not annoying the shit out of us users is directly related to Torvalds’ propensity to tell people “your code is GARBAGE”, instead of sugar-coating it. And that free + open source projects where project leaders don’t do this tend to be crappier. (Does anyone here know a good way to falsify this hypothesis?)
It’s half this, and half an explicit policy “we do not break user space”. Together it meant that if you did anything that screwed up the user space you got told about it at length.
Now Linux culture is established enough that it only really needs the policy, and not the cussing people out to enforce it.
Famous email about it here: https://linuxreviews.org/WE_DO_NOT_BREAK_USERSPACE
I wasn’t aware of that email, only the quote itself.
…not gonna lie, I think that it was beautiful. I have my bones to pick with pulseaudio but come on, you don’t shift blame like this, the guy deserved some smacking.
On-topic: I have my doubts if policy is enough to enforce it, or at least to enforce it in an efficient way.
Well this is it. What really enforces the policy is rejecting commits that break user space.
Now if you’ve got a large enough group of devs, rejecting commits is fine, but if you’ve only got a small group you need everyone to be working productively, and you can see why Linus ended up giving angry feedback about commits that were wasting everyone’s time.