Zig vs Rust. Which one is going to be future?
I think about pros and cons and what to choose for the second (modern) language in addition to C.
Zig vs Rust. Which one is going to be future?
I think about pros and cons and what to choose for the second (modern) language in addition to C.
Honestly C is the future. I don’t know why people would move from C to any other language. It does the job well enough that there’s no reason not to use it.
Think about it. Every modern application depends on a piece of code written in C, not Rust or Zig or any other language (except assembly). It can be used to solve any problem, and works in more places than any other language.
These arguments about “security” and “memory safety” are all pointless anyway in the face of modern code scanning tools. Cross-platform dev can be done trivially with preprocessors. If that’s not enough, I don’t know what to say. Get better at writing C obviously.
Lifetimes and UB should all be kept in mind at all times. You can explicitly mark lifetimes in your C code if you want using comments. Any index-out-of-bounds bugs, use-after-free, etc are just signs that your team needs more training and better code scanning utils. Write more tests!
Anything more complex than a simple
typedefis just a sign that you’re over-engineering your solution.#includeis both simple, and does exactly what you’d expect any reasonable language to do - paste your referenced code inline. It’s genius, and doesn’t require any complicated explanations on namespaces and classes and subclasses and so on.So which will be the future? C obviously.
/s
You got me in the first 3 quarters, not gonna lie!
The number of people that genuinely believe this ( I saw the /s) … Tells me that they haven’t written any useful C or C++ code
… or worked in a large team with juniors and members coming and going over a long period of time.
Or worked on a similar team where the C & C++ was mostly written over a decade ago by dudes in another country who loved multi threading, and some of the “new” features were half-completed about 5 years ago, and nothing is documented, and oh yeah not a single person who did any of that still works at the company. Team is made of great people but all have been here for 0-3 years.
The idea of Rust being roughly as fast and low level as C++ but with improvements to memory safety and concurrency sounds heavenly. I know it’s in the back of most of our minds to look into it for the next big project.
I’m not going to say that C is unusable by any means (and I’m not saying you are saying that). It’s a perfectly usable language. I do think that more people would benefit from exploring other options though. Programming languages are tools, not sports teams. People should familiarize themselves with many tools so they always have a good tool to use for any job.
I think a lot of people believe this because there is some truth to parts of it. I think we see languages like Rust and Zig (and others) popping up to try and solve specific problems better than others.
As for OP’s post, there is no single “C successor” or anything like that. People will use the best tool they know of for the job whether that’s C, Rust, C++, Zig, Python, C#, etc. Many languages will “replace” C in some projects, and at the same time, C will replace other languages in some projects (likely to a lesser extent though).
(Not /s this time)
Oh completely. C is here to stay, C has surpassed language and become protocol cause of libc being so centric to Unix like languages. But it needs to be done carefully and thoughtfully. The other languages are solving some of the pain points C has which I think a lot of people would be better off using than C.
Down that path C may become somewhat of an intermediate representation language for binary interfaces. No one would write it by hand, and maybe for the better
@sukhmel@programming.dev
@programming@programming.dev @modev@snac.bsd.cafe @TehPers@beehaw.org @slacktoid@lemmy.ml C is a terrible binary interface. It doesn’t say who owns a pointer, or what to call to destroy it. It doesn’t normally provide buffer length (even when it does there is no reason to trust it)