Hacker News Reader

Comment by rienbdj | Hacker News Reader

I don't understand the Reason hype when F# / Fable already offers this

As an outsider. F# seems not to get the full blessing. ReasonML is a big thing at Facebook. F# kind of keeps you on the Microsoft stack. Reason has both OCaml based native and JS based interpreted. Reason integrates very well with JS-land. React bindings are maintained by FB itself. Bit-by-bit transitions from JS to Reason are feasible.

Fable, I had to take a look again, but it seems a lot more mature then I remember it was a few years back.

To be honest, ReasonML doesn't seem to have not even half of the community than F# has, nor the IDE tooling for that matter.

Reason is new. F# is not

It thought it was supposed to be a big thing.

I don't know why someone would think that, Reason is actually pretty niche ;-)

Of course, I would love to see it go mainstream, but realistically speaking that will need to happen slowly.

Would say that keeping you on the Microsoft Stack vs the ReasonML/OCAml stack would be more often an asset than a liability especially for most corporations? The .NET Core stack has a better story than OCAml in terms of popularity, library ecosystem, tooling (VS Code/Rider/VS), technology/cross-platform support and general adoption from what I can see. Fable these days is not too bad either from the guide I recently saw online in building applications with it although haven't tried it myself (https://zaid-ajaj.github.io/the-elmish-book/#/).

Companies (like banks, enterprises) still often demand, especially for backends, specific language/runtimes and that tech often translates to Java/JVM or C#/.NET. It is getting to be less important, but we definitely encounter it a lot still and you really can suggest there that you ship something built with ‘ReasonML’ (what is a ‘reasonmul?’) (or ocaml); F#/Scala/Clojure you can get away with especially if you ship closed source (which is still also quite normal in those environments).

From what I've heard a lot of specialized finance applications run OCaml/Haskell/etc. because of the extreme need for safe and reliable code while also preserving performance.

I think 'a lot' is maybe a bit an exaggeration; also, besides maybe Jane Street, it is usually only a small portion of the IT department. But I love to hear more as I am in finance (only encountering Java/.NET basically).

Well for example, one of the most interesting applications of OCaml in finance (well, fintech) is implemented in OCaml and open source: https://gitlab.com/tezos/tezos

Then there's Bloomberg: https://ocaml.org/learn/companies.html

> Bloomberg employs OCaml in a advanced financial derivatives risk management application

Press release from Bloomberg announcing BuckleScript: https://www.bloomberg.com/company/press/open-source-at-bloom...

> We are not currently using BuckleScript on the Terminal, but stay tuned…

This seems to suggest that they were looking at programming Bloomberg terminal applications in BuckleScript, compiled to JavaScript, at least at one point.

There's also LexiFi: https://www.quora.com/Is-the-OCaml-language-used-in-quantita...

I think if you go looking, it's not hard to find OCaml in finance in certain niches.

I looked at the FAQ recently and it says for server side one should compile to JS rather than compiling to native:

https://reasonml.github.io/docs/en/faq

My main interest in reasonml was writing native code. Looks like it will take some time to get to the same level as the JS backend?

Yes, it will, especially in terms of ecosystem and integrations with various other technologies in your stack. But in the meantime I would recommend experimenting with it just because when you hit the sweet spot, it can be very productive. One option is a framework I'm working on: https://github.com/yawaramin/re-web/

F# is not the same thing. It's missing some of the best features of OCaml, e.g. the module system. And no, the object stuff they put into it is not a suitable replacement.

Happy Fable user here.

I think it´s easy to find some feature of a language (such as modules) and claim that every language without it is not viable. You could make the same claim about OCaml not having some of F#s best features, such as operator overloading, inline and computation expressions. But this misses the point. If you´re writing a functional-first web application, then the F# solution and the ReasonML solutions are going to look very similar.

Regarding server run-times and Microsoft lock-in, this is no longer an issue. DotNET Core is open-source and cross-platform. You can also use Fable to compile F# code for Node.

I'm not claiming F# is not viable without the module system, I'm saying F# is not a drop-in replacement for OCaml. There are good reasons people might prefer OCaml. I that that's a reasonable response to this comment:

> I don't understand the Reason hype when F# / Fable already offers this

My only point is that there are good reasons people would prefer ReasonML/OCaml over F#. I never claimed there were no reasons other people might prefer F#.

> DotNET Core is open-source and cross-platform.

The only open-source debugger available for .Net Core is a poorly implemented hack by Samsung, because the vendor-provided debugger was never open-sourced.

https://blog.lextudio.com/the-rough-history-of-net-core-debu...

Reply on news.ycombinator.com