3 ms·
Hmm, have you ever finished any real world scale product in it? All those things you mentioned are mostly toys that mostly inexperienced programmers are only ex
by rexdexpl 6y ago
Hmm, have you ever finished any real world scale product in it? All those things you mentioned are mostly toys that mostly inexperienced programmers are only excited with.
RUST lacks in real world scalability and applicability, it's still toy language even after years of development.
- steveklabnik 6y agoThe foundation member companies are some of the largest technology companies in the world, and are using Rust at scale.
- lights0123 6y agoCritical parts of AWS's Nitro are written in Rust, which powers all of their new high-performance computing—is that a "toy"? What about Dropbox's sync algorithm? You know, the one that runs on every computer with Dropbox installed that is their core business logic, powering file storage for more than 0.5B users? Firefox's Stylo and other parts? They still have hundreds of millions of users.
- pca006132 6y agoI think there are quite a lot of real world scale projects written in Rust... At least for me as an embedded developer, Rust is really superior to C/C++. Although the ecosystem is not that mature and lacks vendor support, it is easier to read/write in Rust and expect the code to work. Do you have any features that you think experienced programmers are excited with? Or features that are required to provide real world scalability and applicability?
- volta83 6y agoI worked on two Rust projects over 500k LOC, one 800k LOC and one that's now at 1.5 million LOC. Number one feature versus Kotlin, C++, C#, C, Haskell, Python, Ruby, etc. is how easy it is to refactor large-scale Rust code bases safely. Refactoring a large C++ code-base is almost impossible to do correctly without introducing lots of bugs, undefined behavior, security vulnerabilities, etc. We did a major refactor (completely parallelized an ~700k LOC application) in a relatively short span of time (6 months) in Rust without issues. We are also not the only ones that did this. Mozilla tried to parallelize Firefox using C++, and this failed so hard, that they actually chose to develop Rust far enough first, to be able to do that in Rust instead, and all that work actually landed on Firefox a couple of years ago already. I wouldn't consider Firefox a "toy" project.