The Linux Kernel is actually hierarchical by design. Anyone can submit a patch, but it then has to go up the maintainer chain to Linus’ final approval before landing mainline, but of course Linus doesn’t review everything himself and implicitly trusts his maintainers.
So part of the Rust drama a few months ago was accusations that despite the stated goal of rustifying some subsystems, the existing hierarchy is sometimes acting in bad faith and unwilling to learn the basics of Rust to talk ABI or generally accommodate the reasonable needs of Rust devs. Asahi Lina had an impressive writeup of her Rust contributions to the Apple Silicon GPU driver and the frequent, demotivating difficulties she had with maintainers refusing to learn anything that isn’t C or to acknowledge errors like race conditions in their C code. Some insanely talented people are being kept at arm’s length by the kernel community over petty turf wars that look very much like a symptom of institutional rot. Which isn’t very surprising to me having met some unrelated but very highly opinionated (and sometimes very confidently incorrect) greybeards of similar ilk.
I don’t have a horse in that race or a solution to the kernel issues, but it’s interesting to watch how at scale even kernel OSS devs fall into the same trappings as any institution with a hierarchy. We’re all just human, and even when working for an organization with the most noble of goals we must keep an eye out for hierarchies and institutions and rules and processes.
The Linux Kernel is actually hierarchical by design. Anyone can submit a patch, but it then has to go up the maintainer chain to Linus’ final approval before landing mainline, but of course Linus doesn’t review everything himself and implicitly trusts his maintainers.
So part of the Rust drama a few months ago was accusations that despite the stated goal of rustifying some subsystems, the existing hierarchy is sometimes acting in bad faith and unwilling to learn the basics of Rust to talk ABI or generally accommodate the reasonable needs of Rust devs. Asahi Lina had an impressive writeup of her Rust contributions to the Apple Silicon GPU driver and the frequent, demotivating difficulties she had with maintainers refusing to learn anything that isn’t C or to acknowledge errors like race conditions in their C code. Some insanely talented people are being kept at arm’s length by the kernel community over petty turf wars that look very much like a symptom of institutional rot. Which isn’t very surprising to me having met some unrelated but very highly opinionated (and sometimes very confidently incorrect) greybeards of similar ilk.
I don’t have a horse in that race or a solution to the kernel issues, but it’s interesting to watch how at scale even kernel OSS devs fall into the same trappings as any institution with a hierarchy. We’re all just human, and even when working for an organization with the most noble of goals we must keep an eye out for hierarchies and institutions and rules and processes.