5 ms·

Criticizing the Rust Language, and Why C/C++ Will Never Die

6 points by pyprism 9y ago

topspin 9y ago
Myopic rant. "If your C++ code is not good enough or Java code is painfully slow, it's not because the technology is bad - it's because you haven't learned how to use it right." More 'program better and you don't need Rust.' This was said of C++ by C programmers and it regresses until we're programming punch cards in octal. At one point the argument is (indirectly) made that Rust is on the "innovative yet too idealistic" side of the ledger. This is bogus; Rust is pragmatic. Otherwise release code would pay the cost of detecting integer overflows and unsafe wouldn't exist. Rust just doesn't cop out on all the stuff that _can_ be dealt with in a reasonable manner. "static and dynamic analyzers [exist for C++]" Dismissing the obvious value of providing one uniform and precise analyzer integrated into the compiler is just bad thinking. "But these are actually two dialects of one language" Unsafe changes a small and finite set of rules. It is an outrageous exaggeration to claim a new 'dialect' is created by unsafe. At one point the following construct rears it's ugly head again. It's become a favorite for Rust haters: Vec<Rc<RefCell<Box<Trait>>>> The cause of this somewhat unergonomic construct was dealt with over two years ago[1], and a workable alternative using Cell<> existed at the time. Also, how does a C++ advocate look at that type and somehow imagine it to be worse that the utterly incomprehensible line noise that C++ templates routinely devolve into? Hell, almost any given pointer to function in C is that bad or worse. "it doesn't have a decent inheritance and exceptions" C++ doesn't have 'decent' inheritance either; most C++ programmers don't even understand how inheritance in C++ actually works. Arguing that the lack of exceptions in Rust is somehow a problem is deeply wrong; better and well defined alternatives are provided. C++ exceptions are a failed language feature and an enormous fount of UB. In all likelihood you're reading this using a browser written by developers that have banned C++ exceptions from their code bases (Google and Mozilla.) "Rust is safe indeed but, unfortunately, far from fast" The one Computer Language Benchmarks Game benchmark cited here is cherry picked. Rust does very well against C++ in that set of benchmarks, winning 6 of 10 at the moment and only a little less well against C. And we all know the sort of issues inherent to those benchmarks; citing them without qualification is either poor thinking or an insult to your intelligence. Two terrible alternatives. "Macros used as a crutch to make up for the excessive verbosity caused by the absence of normal exceptions" That doesn't make sense. mic drop [1]https://github.com/rust-lang/rust/pull/24619 https://github.com/rust-lang/rust/pull/24619
- igouy 9y ago
  > Myopic rant. A 2 year old blog post adequately discussed back then -- https://news.ycombinator.com/item?id=9531822 https://news.ycombinator.com/item?id=9531822 > And we all know the sort of issues inherent to those benchmarks Inherent to those benchmarks or inherent to benchmarks? Innuendo.

6 points by pyprism 9y ago

Myopic rant. "If your C++ code is not good enough or Java code is painfully slow, it's not because the technology is bad - it's because you haven't learned how to use it right." More 'program better and you don't need Rust.' This was said of C++ by C programmers and it regresses until we're programming punch cards in octal. At one point the argument is (indirectly) made that Rust is on the "innovative yet too idealistic" side of the ledger. This is bogus; Rust is pragmatic. Otherwise release code would pay the cost of detecting integer overflows and unsafe wouldn't exist. Rust just doesn't cop out on all the stuff that _can_ be dealt with in a reasonable manner. "static and dynamic analyzers [exist for C++]" Dismissing the obvious value of providing one uniform and precise analyzer integrated into the compiler is just bad thinking. "But these are actually two dialects of one language" Unsafe changes a small and finite set of rules. It is an outrageous exaggeration to claim a new 'dialect' is created by unsafe. At one point the following construct rears it's ugly head again. It's become a favorite for Rust haters: Vec<Rc<RefCell<Box<Trait>>>> The cause of this somewhat unergonomic construct was dealt with over two years ago[1], and a workable alternative using Cell<> existed at the time. Also, how does a C++ advocate look at that type and somehow imagine it to be worse that the utterly incomprehensible line noise that C++ templates routinely devolve into? Hell, almost any given pointer to function in C is that bad or worse. "it doesn't have a decent inheritance and exceptions" C++ doesn't have 'decent' inheritance either; most C++ programmers don't even understand how inheritance in C++ actually works. Arguing that the lack of exceptions in Rust is somehow a problem is deeply wrong; better and well defined alternatives are provided. C++ exceptions are a failed language feature and an enormous fount of UB. In all likelihood you're reading this using a browser written by developers that have banned C++ exceptions from their code bases (Google and Mozilla.) "Rust is safe indeed but, unfortunately, far from fast" The one Computer Language Benchmarks Game benchmark cited here is cherry picked. Rust does very well against C++ in that set of benchmarks, winning 6 of 10 at the moment and only a little less well against C. And we all know the sort of issues inherent to those benchmarks; citing them without qualification is either poor thinking or an insult to your intelligence. Two terrible alternatives. "Macros used as a crutch to make up for the excessive verbosity caused by the absence of normal exceptions" That doesn't make sense. mic drop [1]https://github.com/rust-lang/rust/pull/24619 https://github.com/rust-lang/rust/pull/24619

igouy 9y ago
> Myopic rant. A 2 year old blog post adequately discussed back then -- https://news.ycombinator.com/item?id=9531822 https://news.ycombinator.com/item?id=9531822 > And we all know the sort of issues inherent to those benchmarks Inherent to those benchmarks or inherent to benchmarks? Innuendo.