Thanks for sharing. This makes a lot of sense to me. I’m not close to kernel development but any change to an established entity impacts the interest of people in it. So a primary obstacle change encounters is that it may diminish power people obtained in the establishment.
These people, while well intentioned in pushback, have difficulty separating benefit to the entity and benefit to themselves. So it takes a lot of work to impact change, and I understand the distance it has from normal technical work.
He did a presentation on making a fs in rust, the response from this dev was "we won't be converting to your religion. We are going to continue modifying c code and your dumb rust code will always be a second class citizen."
Literally just hostility for hostility sake. He wasn't proposing to rewrite everything in rust. He wasn't proposing that everyone learns rust. He was just giving a freaking presentation on making a rust fs that integrates with the kernel and was immediately trashed for spreading religion or whatever.
What was literally said didn't include the words "always" or "your dumb rust code". From the subtitles with slight manual fixes:
> I suspect part of the problem here is you're trying to convince everyone to switch over to the religion as promulgated by rust and the reality is that ain't going to happen because we have 50 plus filesystems in Linux, they will not all be instantaneously converted over to rust. Before that happens we will continue to refactor code because we want to make the C code better. If it breaks rust bindings, at least for the foreseeable future the rust bindings are a second class citizen and those file systems that depend on the rust bindings will break and that is the rust bindings problem, not the file system community at large's problem and that's going to be true for a long long time, okay, and I think we just simply need to accept that
That seems rather technical to me. To put words into their mouth because they had a snarky tone, or claim the presenter was "immediately stacked for spreading religion or whatever" hardly is.
They literally said "spreading your religion" at the very start of their critique. I think I'm correctly reading the tone of the rest of the comment even if I didn't verbatim quote them.
From the closed captions with minor editing. In response to "we are going to provide API bindings"
> I suspect the problem here is you are trying to convert everyone to the religion as promulgated by rust and the reality is that ain't gonna happen. Because we have 50 plus filesystems in linux and they will not all be instantaneously converted over to rust. Before that happens we will continue to refactor code because we want to make C code better. If that breaks the rust bindings at least for the foreseeable future the rust bindings are a second class citizen. Those file systems that depend on the rust bindings will break and that is the rust bindings problem not the file system community at large problem. And that's going to be true for a long long time, OK. And I think we just simply accept that right because the answer you are not allowed to refactor C code because it would break 5 critical file systems that distros depend on is like not a star, OK.
> They literally said "spreading your religion" at the very start of their critique.
And that critique is near the end of all the people who asked questions or made comments. So no, the presenter wasn't "immediately stacked" when making that presentation.
Linus has the wisdom to see the writing on the wall - the kernel is becoming too
complex to be maintained and extended by newcomers; the pool of C ninjas that can handle it reliably is diminishing. This kind of foresight is what makes Linux trustworthy in the first place. A lot of current individual contributors lack that vision and fail to understand the reason for introducing what they see as unneeded complexity.
The other reason Linux is trustworthy because of the technical meritocracy culture. But since Rust is not a technical choice, as Rust cannot per-se do things that are impossible in C - you can do anything in C - the introduction of Rust falls outside the parameters of the regular rough-but-mostly-fair Linux exchanges. This leading to toxic reactions from the less socially-abled.
> But since Rust is not a technical choice, as Rust cannot per-se do things that are impossible in C - you can do anything in C
Have anything in particular in mind? It's been my understanding that Rust actually can do everything C can, but the additional safety guarantees imposed by the type system make it a bit harder.
Yup, anyone defending this behavior needs a screw tightened.
The dev "pushing back" was being an asshole probably because of deep insecurity of gasp there being other programming languages than C.
Literally rust is currently being proposed for maybe new drivers and basically nothing else ATM. Nobody is suggesting rewriting code in rust or forcing people to learn rust. It is literally just "Hey, might be nice to write this driver for my mouse in a memory safe language".
It's inevitable in any organisations when dealing with change. Identifying resistance and coming up with strategies to overcome is a key part of change management. Telling people their reaction is "kneejerk pushback" usually does not help ;)
It's probably more difficult to handle in open source than in a company because authority less powerful and it's more difficult to wave both carrots and sticks...
At a talk presenting a personal project? Do you also feel it's OK to go to talks on Java projects to make statements like "Well, why didn't you write this in Go? It's a better language that serious developers are using and nobody will want to contribute to your dumb project using java"
