Ask HN: Elm, would you use it in production?
I'm a team of one creating apps for internal use. I want to use a tool I'll enjoy. Elm seems great but is it wise to use something so niche in production code?
- Componentized design is eschewed by the community, so it's very hard to break down apps into small easily-combined pieces.
- The learning curve for new engineers is relatively high, and drops off pretty significantly after a month or so. It might be easier if your team is already familiar with strongly typed languages, but there's quite a bit to learn to get up to speed.
- The lack of a roadmap is the only true non-starter. Our team ran into a compiler bug, and the options were either to patch it ourselves (which is unlikely since no one on our team knows Haskell or has that much free time), or to wait an indeterminate amount of time and pray it gets patched. If you look at the release history, you'll see that updates to the language are infrequent.
I would still highly recommend Elm for a side project, just not for something that a business or team relies on.
However maintaining the code is a concern. It’s pre version 1.0 and each release has breaking changes.
The general direction is away from supporting the whole web api, and making it harder to write your own ffi to work around this as you have to use ports as they have scrapped other methods in user space like native code.