Hacker News Reader

Comment by platinumrad | Hacker News Reader

Maybe not the right place to ask, but "Better standard library" appears in the section detailing future improvements that ReasonML might bring to the OCaml ecosystem. How do they plan on doing this, given that the major standard library camps seem relatively entrenched in OCaml today? Which one does the ReasonML community favor/intend to contribute to?

View on HN

Base[1] is a good candidate for an OCaml standard library. It's got tons of functionality, results in binaries with small footprints, is cross-platform, and has a fairly stable API. It's also code that has been and continues to be battle tested every day in production systems at Jane Street. I'm biased, as I'm one of the maintainers, but trust me it's good, it's good.

[1]: https://github.com/janestreet/base

How does Base compare to Core?

https://github.com/janestreet/core

Seems like Core depends on Base (https://github.com/janestreet/core/blob/master/core.opam)

Core depends on Base. The parts where they differ include: more data structure, and bindings for system calls and other functionality available from libc.

There's also core_kernel, which is core but without the libc support.

We also just released improved docs for Base. Still more to do, but much better than before:

https://ocaml.janestreet.com/ocaml-core/latest/doc/base/

I'm not convinced that Base will solve our problems. It's not, afaik, more portable, more lightweight, more performant, or more feature-full than [containers](https://github.com/c-cube/ocaml-containers) (which has reached 1.0 and also does a bit of IO and Unix, as a stdlib extension). In addition to that it's not compatible with programs that use the stdlib right now, so the fragmentation will stay for a long time (unlike for build systems where you "only" have to change the build scripts, not the code itself).

(Of course I'm biased too, I wrote containers ^^)

Reply on news.ycombinator.com