Recorded live on twitch, GET IN
### My Stream
### Best Way To Support Me
Become a backend engineer. Its my favorite site
This is also the best way to support me is to support yourself becoming a better backend engineer.
MY MAIN YT CHANNEL: Has well edited engineering videos
Discord
Have something for me to read or react to?:
Kinesis Advantage 360:
Get production ready SQLite with Turso:
[ad_2]
source
bikeshedding engineer lessgoo
but i agree though.. rust is good but so much so it solve problems that are never exist.
"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.” (?)
IDK, Linux kernel developers are real clowns. Rust for Windows mostly is just autogenerated bindings. 5 million lines of autogenerated bindings code. That's it.
it feels like the rust community are victimizing themselves a lot when in reality the C guys could be worried about something else, backward compatibility?! the rust guys should go build a new OS, maybe.
Rust does not have a stable abi whole C does (all be it a very anoying one)
Now not the biggest deal u can just do system v abi in rust. It is ironically hard to pass rust constructs along the call boundary but for a Unix clone that means jack shit
Existing maintenance in the Linux kernel must cease immediately
All new code must use the latest version of React
All APIs shall be replaced with http endpoints that return ambiguous blobs of JSON
Oh my god, the moment someone invokes "Religion of" in a tech talk, you know they're not worth listening to.
having more languages in the source code increases complexity, is it worth it after all? when bugs start to appear, will there be someone to solve them? or will it just be another problem to solve?
If these guys are doing this for free, they probably have no interested in learning a new language. C is a tiny language, there isn't much to learn. Rust looks decent, but far from trivial.
Its the rust guys who bring drama here to be honest… They are a sect… All the guy says is that encoding these into type system makes refactors very hard and changes very hard. He says he will not care as a C dev for the rust bindings and if the bindings go wrong or out of sync for this. That is all. It is reasonable. It is a C system where Rust guys enter… They however expect the C guys to tell them changes – thus generate a bunch of extra work for the C guys who would otherwise just chill or fix other real issues.
The problem is that the rust people want to rewrite these just so it is in rust to be honest… And while doing so expect the C people to spend resources on THEM. Why? If it breaks go and fix it yourself… what's so wrong?
gaaahhh.. omg , how many tiimes i told people that no compiler will help you avoid mistakes. "my code is correct, so i won't be yanked back to it again" (because we all know, your code smells, and redgardles if it correct from standpoint of compiler and "types" – its doing wrong things! surprise, MF.. "correct" code can do wrong thing!!
what a brainless people get into kernel development, who don't even understand such basic things?
so you need haskell
Looks like C guys took pickets from the Chesterton’s fence and beat Rust guys senseless
I've seen some similar friction when it comes to automated trading as what seems to be happening between Rust and C developers in the kernel. The only way I've observed to succeed in introducing something entirely new to something already well established is to prove the new thing's worth to key decision makers.
For example, trying to prove the worth of automation to many people comfortable with the status quo can be threatening. For some traders they may see automation as something that will take away their job. The C developers likely see Rust as a risk to them continuing to program what they are interested in using C. Effectively, they have a personal reason to not have Rust enter the kernel. However, this is not the same of key decision makers whose primary interest is to do whatever is good for the kernel itself.
So, perhaps what is best for the Rust team is to work with these key kernel decision makers (the lords of the fiefdoms) to see how they can start improving things at a small scale until Rust has proven itself to take on more critical parts of the system. You can't make everyone happy, but it is a lot easier to make more people happy if you are solving headaches many people are currently having.
Honestly I'm all for the Rust takeover – the only issue I have with it personally is how it looks and feels to me (with a preference towards C-like languages – and I'm not a fan of ulltra-abbreviated keywords).
It's not a logical argument, but I feel like the shift to Rust is difficult because we're asking people to make two changes: Paradigm shift in terms of how they think about flow/structure + Code style and syntax reeducation.
Coming from C, rust may "look ugly" to some. An intermediate "Rusty C" or something that functions like Rust but maintains the syntax stylings of C (with added/removed keywords etc) would make the transition more palpable (and potentially translatable to Rust).
What follows might be a garbage take by me, but it doesn't fit anywhere else:
My hope was that languages by now would have become abstracted enough where the syntax is not married to the paradigm and operations enforced by the compiler, so that you could use any syntax styling you wanted as long as it conformed to the underlying operations which would be an intermediate before assembly; but unfortunately the idea of language abstraction didn't really catch on when Python and Java abstracted to bytecode – there was no huge plethora of languages using it that stuck around – and as much as it pains me, C# and VB.NET had more success with language abstraction considering MSIL. This is a human and social problem as much as a technical one; you need adoption for the solution to work.
Because of this we're still stuck in a world with the tough decision of C vs Rust, of being forced to use Javascript/WASM as the compile target for web applications, and many other area-specific languages.
Lol, the comment about bike shedding made me laugh. I feel like if we don't bike shed something we feel we aren't doing our job and it makes us look so clever in front of non techies 😂
Kernel doesn't do assertions cos it assertion is crash and you don't want to crash in kernel.
Even though Rust does not guarantee correctness you can encode various non memory safety related properties into the type system, which helps correctness. Rust model of aliasing XOR mutability also lends itself very well to formal verification, which can guarantee correctness. There are a lot of projects on formal verification with Rust. Getting formal verification to work at scale is much more difficult in C.
linux is C, if rust is better, then make make a new OS and prove everybody who disagrees wrong
this feels like hollywood when they start making 10 sequels of terminator, because they know better and everything is better now, yet none of these movies are worth watching and only try to be relevant because they ride on the back of something made before, that was actually good