Honestly, it’s sad to see OCaml being replaced by Rust (or being discarded in favor of Rust, like it happens for other projects when they’re choosing a stack), mainly due to popularity.
I understand that Rust has its uses (and low-level systems programming is definitely one of them), but the same area is also very well covered by OCaml (only need to look at MirageOS for proof), so it really ends up being mostly about popularity, in some cases.
I wonder if OCaml will ever be able to overcome this problem.
I think it’s opposite: Rust is wildly more popular than OCaml because it happens to be just better across many axis relate to practical programming. Relative to OCaml, it has:
one standard library
one build tool
windows as first-class platform
stable interface for syntax extensions
access to OS threads
It really is a shame that we don’t have an ML implementation which has “all the nice things”, I feel like a cluster of applications which is written today in Go/Java/C# and, lately, Rust, would’ve been simpler in some vague “modern ML”.
Yeah, and a lot of these things are not really about the actual language at all, but rather the ecosystem around it. (The language has influence on things like stdlib and OS threads of course, but still.) My heart will always have a soft spot for OCaml, but the ecosystem is a bit infuriating.
so it really ends up being mostly about popularity, in some cases.
But that is a fair reason right? It’s far easier to attract new contributors if a project is in Rust than in OCaml. There simply are many more Rust programmers (I don’t think this needs a citation anymore). Running a large project is all about making compromises and sometimes that means giving something up that may like more for something that is better in practice. Besides that, the choice could be worse, Rust the compiler has a history in OCaml and Rust the language is also very clearly influenced by ML/OCaml.
Honestly, it’s sad to see OCaml being replaced by Rust (or being discarded in favor of Rust, like it happens for other projects when they’re choosing a stack), mainly due to popularity.
I understand that Rust has its uses (and low-level systems programming is definitely one of them), but the same area is also very well covered by OCaml (only need to look at MirageOS for proof), so it really ends up being mostly about popularity, in some cases.
I wonder if OCaml will ever be able to overcome this problem.
I think it’s opposite: Rust is wildly more popular than OCaml because it happens to be just better across many axis relate to practical programming. Relative to OCaml, it has:
It really is a shame that we don’t have an ML implementation which has “all the nice things”, I feel like a cluster of applications which is written today in Go/Java/C# and, lately, Rust, would’ve been simpler in some vague “modern ML”.
Yeah, and a lot of these things are not really about the actual language at all, but rather the ecosystem around it. (The language has influence on things like stdlib and OS threads of course, but still.) My heart will always have a soft spot for OCaml, but the ecosystem is a bit infuriating.
But that is a fair reason right? It’s far easier to attract new contributors if a project is in Rust than in OCaml. There simply are many more Rust programmers (I don’t think this needs a citation anymore). Running a large project is all about making compromises and sometimes that means giving something up that may like more for something that is better in practice. Besides that, the choice could be worse, Rust the compiler has a history in OCaml and Rust the language is also very clearly influenced by ML/OCaml.
Why lie about this when it takes reading the article for two minutes to know it’s not true?