It’s fair to want to learn (and it’s certainly a good skill to have), but the question is what you’d rather see in a large, production environment. Guard rails are usually there for a reason. As for the control: you actually can program memory-unsafe (and in kernel development you often have to!) in Rust. The difference is that in Rust it’s explicitly marked by an unsafe block:
unsafe {
...
}
That way you get the same, fine-grained control over low-level processes, but someone else reading your code can at a glace spot where potential memory bugs may be.
In the end, languages are a tool. Especially for personal projects, everyone should just go with what’s fun to them. I personally think it makes sense, logistically, to slowly transition legacy C-based projects to Rust, because it makes onboarding new developers easier, while keeping the same memory safety that requires years of experience otherwise, basically for free. But there’s really no rush to rewrite anything that’s working well in Rust
Alas, when there is no difference between unsafe wrapper in Rust and C, then why learning Rust, if one wants to go for managing the memory manually? Especially when considering the complex way of coding in Rust?
Another problem: going the easy way and forgetting the tricky parts - if Rust allows for unsafe code, but it is safer to put it into a « safe » mode, so why I need to take the burden and deal with unsafe code? This will evidently lead to the situation that less and less unsafe blocks will be used, which finally leads to a situation where the programmer forgets the in and outs of manual memory management. You can see it as the principal aim of writing memory safe code, but to me it is also a way of « delearning » by learning. I see here the reason why so many young programmers are opting for Rust, because manually managing the memory in larger projects like in C is a question of knowledge and experience which does not come in one day. I also doubt that following just the compiler is a good approach.
I agree totally with your last points though! Coding should mean to have fun and be the same time mentally challenged due to complex algorithms or demand for better code in general.
It’s fair to want to learn (and it’s certainly a good skill to have), but the question is what you’d rather see in a large, production environment. Guard rails are usually there for a reason. As for the control: you actually can program memory-unsafe (and in kernel development you often have to!) in Rust. The difference is that in Rust it’s explicitly marked by an unsafe block:
unsafe { ... }That way you get the same, fine-grained control over low-level processes, but someone else reading your code can at a glace spot where potential memory bugs may be.
In the end, languages are a tool. Especially for personal projects, everyone should just go with what’s fun to them. I personally think it makes sense, logistically, to slowly transition legacy C-based projects to Rust, because it makes onboarding new developers easier, while keeping the same memory safety that requires years of experience otherwise, basically for free. But there’s really no rush to rewrite anything that’s working well in Rust
The sound bite I heard was “the unsafe keyword makes memory bugs greppable.”
Alas, when there is no difference between unsafe wrapper in Rust and C, then why learning Rust, if one wants to go for managing the memory manually? Especially when considering the complex way of coding in Rust? Another problem: going the easy way and forgetting the tricky parts - if Rust allows for unsafe code, but it is safer to put it into a « safe » mode, so why I need to take the burden and deal with unsafe code? This will evidently lead to the situation that less and less unsafe blocks will be used, which finally leads to a situation where the programmer forgets the in and outs of manual memory management. You can see it as the principal aim of writing memory safe code, but to me it is also a way of « delearning » by learning. I see here the reason why so many young programmers are opting for Rust, because manually managing the memory in larger projects like in C is a question of knowledge and experience which does not come in one day. I also doubt that following just the compiler is a good approach. I agree totally with your last points though! Coding should mean to have fun and be the same time mentally challenged due to complex algorithms or demand for better code in general.