Fusion Programming Language

by efrecon • 29 days ago

113 48 29 days ago

efrecon• 29 days ago

Fusion is a programming language designed for implementing reusable components (libraries) for C, C++, C#, D, Java, JavaScript, Python, Swift, TypeScript and OpenCL C, all from single codebase.

hav0c• 29 days ago

Reminds me of https://xkcd.com/927/

jdonaldson• 26 days ago

I'd love to see a comparison to Haxe. https://haxe.org/

I wonder what performance and generated code size/quality look like.

nine_k3• 26 days ago

Apparently Haxe cannot target OpenCL. It can target PHP and Lua instead.

jp3141• 26 days ago

From https://github.com/fusionlanguage/fut/discussions/119#discus...

I've read about Haxe in 2000s, before I started my work on Fusion. They have different design goals: in Haxe you create whole apps, in Fusion you create components to be used from other languages. Haxe has syntax similar to the (now dead) ActionScript, Fusion is similar to C#. Fusion transpiles to C, D, Swift, TypeScript, OpenCL and Haxe does not.

broken-kebab• 26 days ago

But what is the goal of Fusion then in pragmatic words? E.g. Haxe helps to save on development costs when targeting multiple platforms with emphasis on graphical interactive apps (games mostly).

raphinou• 26 days ago

At a time I was very interested in haxe, but their focus on games made it (perceived as) lacking in the area I wanted to use it (cross platform client-server apps). Recently rust seems to have taken this role for me.

jdw64• 26 days ago

Wow, this is really fusion. I like it.

knhung• 26 days ago

The idea it's good but hard to make it good. A universal language is hard to optimise for a particular language.

nine_k3• 26 days ago

The point, AFAICT, is not in using all capabilities of all the target languages. Rather, it's about expressing some narrower class of computations and grafting them seamlessly into the target languages. Think of data formats, parsers, network protocols, stuff like handling and rendering of text, etc.

g42gregory• 26 days ago

What are the use cases? I am curious why Rust was not targeted.

nine_k3• 26 days ago

I suppose you can write various algorithms in it, and have that code natively trsnspiled to different languages, for ease of native interoperability. It's unlikely to produce the absolutely most optimized code, but the lack of the interface translation barrier (aka FFI) may more than compensate for it.

Rust is not easy to target efficiently, due to the borrow checker, and they likely don't want to dyn Box everything.

NonHyloMorph2• 25 days ago

I first thought you were going to ask about the usecases for the austrian fut lol

Panzerschrek2• 26 days ago

implementing reusable components (libraries) for C, C++, C#, D, Java, JavaScript, Python, Swift, TypeScript and OpenCL C, all from single codebase

Why is this needed? I can't imagine that. I am sure writing code in fusion will produce C++ and Python code which is suboptimal and doesn't fit well in these languages.

4k0hz2• 26 days ago

I think the target application is writing the same algorithm in multiple places with a guarantee that the logic will be based on a single source of truth. Not unlike Protocol Buffers work to standardize data layout across platforms.

It still feels overcomplicated compared to the standard solution of writing a library in a compiled language you like, exposing a C ABI compatible interface, and hooking it up to any language that can work with that (i.e. any language).

BobbyJo• 26 days ago

C code called from other languages has terrible ergonomics. Writing it sucks, debugging it sucks, maintaining it sucks.

I don't know if fusion is the solution, but I know C isn't.

zigzag3123• 26 days ago

What about a better ABI abstraction?

imtringued2• 26 days ago

You don't need a C ABI here because the language could in theory generate both the library and the cross language bindings, the underlying mechanism then becomes a mere implementation detail.

I feel like there is a huge missed opportunity that Fusion Lang overlooked.

Panzerschrek2• 26 days ago

writing the same algorithm in multiple places with

This seems for me to be reasonable. Like for a video game with server written in Go/C++/Rust and client written in C#/Javascript, where the same calculations should be done on both sides.

astrobe_• 26 days ago

ORMs and variations like Protobuf or things that have to be cross-plateform in the wide sense. The perspective that the same source will behave the same in various environments, and "velocity" trumps performance considerations. If you want to work on things where performance matters, consider embedded/firmware programming ;-)

BSTRhino• 26 days ago

I once had a C# server with a TypeScript frontend (this was for work, so I didn't choose the tech stack). We had one part of the code which had a lot of calculations. We wrote the code in a C# library, then used Roslyn to analyse the C# code and generate TypeScript code that matched it identically. This way the client could visualise the results on the client with zero latency, no server roundtrip, but ultimately the server would be in charge of the final calculations. It was a great solution, I would do it again.

We could have potentially used Fusion for this had it existed, but I don't know if we would. We didn't have to support 12 different languages, only 2.

Twirrim• 25 days ago

In theory having it in the same language could provide performance benefits, vs reaching out to a C library and dealing with C interop fun. Any language that is JIT'd would have native bytecode to optimise instead of a black box.

hyperhello• 26 days ago

The syntax is so warm and simple, but not too simple. I like it, but I wonder how debuggable it can be, or if there is an interpreter for learning.

teo_zero• 26 days ago

Nice idea. I'm just wondering how to debug code written in fusion... probably you must focus on one of outputs, debug that one, and then back-fit the changes to the fusion source. :/

speps• 26 days ago

It’s not a problem in practice if debug data is emitted properly. For example I used the VS Code C++ debugger extension with Zig.

anonzzzies• 26 days ago

As others say ; how does debugging work? I was a user of Haxe and Monkey and both had/have serious debugging issues; I think this can only work if it has an interpreter for itself (self hosted I guess ideally) with a proper debugger.

eoanermine• 26 days ago

Nice stuff, I'm working on something like Kaitai Struct, and it'll be very convenient to codegen to Fusion, to support multiple target languages.

However, what I like the least in all kind of software is underdocumentation. There's some reference, it's not bad, but actually it's just bunch of examples. If I want to find out exactly how a particular type or statement translates into a specific language, I'll have to dig into the code.

FattiMei• 26 days ago

How (and why) are you targeting OpenCL? It has quite a different programming model

internet_points• 26 days ago

but not Haskell, guess they gave up on picking a string type =P

pelasaco• 26 days ago

I like it. I just wonder if today transpiling is better then Agentic Programming. Any specific use case where i should use it? With a good model, SKILLS.md, AGENTS.md and prompts I am able to do - allegedly - a better job then a transpiler, no?

poppadom1982• 26 days ago

It's a nice idea but the C portion expects an array of unknown size, no bounds checking possible at all and it's also not returned how large the PNM image is. I also don't understand why glib is needed here to get the array to work. It's also missing the stdint.h include to be able to use uint8_t.

gr4vityWall2• 26 days ago

It seems similar in concept to Haxe, but focused on writing libraries which are then called from the target language, rather than building your whole application in Fusion. Is that correct?

How are exceptions translated to languages which don't have exceptions, like C?

The amount of targets seems impressive, but I don't see anything resembling a standard library. Usually, it's in the standard library where diverging behavior between targets shows up. Haxe's documentation tends to cover those issues, and lets you know when a behavior is target-dependent, or unspecified. They do try to be consistent between targets whenever possible in a performant way.

Loreline[0] is a recent project which leverages Haxe to build a cross-platform library. Is there a similar project showcasing how Fusion can be used for that?

[0] https://loreline.app/en/docs/

harrisi• 25 days ago

https://github.com/fusionlanguage/fut/blob/master/doc/refere...

gr4vityWall2• 23 days ago

Thanks for the link. Seems like it isn't ready for general usage yet.

zabzonk• 26 days ago

We seem to be regressing to the overuse of classes again.

xiaoyu2006• 26 days ago

Bad idea. It either takes a useless subset of all languages or do weird abstraction that wouldn't fit in a language otherwise.

hathym• 26 days ago

Now we need an UnFusion programming language that can take code written in any language and convert it into Fusion. With that, you could transform any project from language A into language B.

crvdgc• 26 days ago

That language should be called Fusion and this language Fission.

NonHyloMorph2• 25 days ago

Nice. LessIsMore

Alifatisk• 26 days ago

I was a bit surprised to see Ds logo there. Cool.

stefs• 25 days ago

somewhat funny to the immature mind: "fut" - the name of the compiler binary - is austrian slang for vagina, but in the most obscene, vulgar way possible. a bit like "cunt", but worse.

andrekandre• 25 days ago

from the example:

  public class HelloFu
  {
    public static string GetMessage()
    {
        return "Hello, world!";
    }
  }

translated into swift this gives a class, which means i'm dealing with reference semantics (and potentially threading issues) and probably not something i necessarily want... i tried struct, but there is no struct keyword in fusion it seems...

Fusion Programming Language

Related Stories