Disallusioned with Julia anyone got thoughts on Carbon? | Hacker News Reader

Disallusioned with Julia anyone got thoughts on Carbon?

5 pointsmrsofty15 comments

Hi all we've examined Julia for a while now and decided it's time to move on. Anyone looked at Carbon? Maybe moving on from Python, ObjectiveC or C++.

View on HN

You should explain what did not work with Julia since there is extreme likelihood the situation will be worse with Carbon, given Carbon is just getting started. The best option might be to stick with Python (or C++)

Carbon is such a werid choice. I don't see anyone in their right mind using Carbon for anything meaningful for a lot of reason. As others mentioned, it's really incomplete and you cannot trust google to not kill it out of nowhere.

I don't know what OPs project looks like, but in our company the choice has a clear path: - performance/low level: Rust, Cpp

- CPU/memory is not the bottleneck and I don't need much concurrency / it's data science, machine learning, ETL ... : Python

- concurrency is key, highly distributed system: elixir, golang, erlang

- programming based on mathematical models: Haskell

If you're going to consider obscure languages, and you liked Julia theoretical features but got disappointed by the implementation and perhaps instability, I suspect you will enjoy Common Lisp.

Check this out: https://lispcookbook.github.io/cl-cookbook/numbers.html

Lisp is fun but not really relevant now we have Julia.

Lisp will always be relevant.

It’s a paradigm.

Is this a personal or business decision you are talking about?

Can you elaborate the cons that made you decide to move away from Python, C++, and Julia?

Julia makes sense to me, as the youngest production language and community.

But Python? It is the #1 most popular programming language globally. C++ is also very popular. This means you can easily find people to hire globally (making business easier and more economical). You can also easily find code doing almost anything you want---speeding up time to value and time to ship.

I'm wondering what cons are driving your decisions to move on from Python and C++?

Carbon seems much more "low-level" and less "dynamic" than Julia? They don't seem to play in the same domain. If you're looking for a lower-level language than Julia, supporting operations on vectors, Zig is interested (usable but still in development): https://ziglang.org/documentation/master/#Vectors. But it relies on manual memory management (no garbage collector).

I don't like Julia either but Carbon isn't ready for use - https://github.com/carbon-language/carbon-lang - and even if it was, it would still need an ecosystem of libraries which it doesn't have.

I'm not sure I understand the premise of the question. What parts of Julia seem to be lacking?

I have a comment on this site about Carbon from its initial announcement: https://news.ycombinator.com/item?id=32158805

A short summary of my feelings is that it has some syntactic oddities, it's too incomplete in some areas to fully evaluate (particularly error handling), and it completely botches overflow. Whatever your opinions on undefined behavior on integer overflow, there really is no justification for keying undefined-versus-well-defined on signed versus unsigned.

Never heard of Carbon before, but I took a look at some code, and it seems like you could change a few symbols and it starts to look a lot like Rust...

Given that Carbon isn't ready, I think if you're porting something from C++ and Carbon looks good to you, Rust might be the way to go.

Alternatively, if you wanted something more functional, I've heard good things about OCaml

I didn't find a solution to read in a .png-file and iterate over the rows and in every row over the columns, is the low level stuff somehow hidden? Found some higher level stuff libraries here and there, but if it's not easy to do some low level operations, the language would be useless for me.

Carbon looks interesting but there's no working compiler yet.

Julia and Carbon fill very different niches. Can you say something about the type of work you’re doing that these would be the two languages you’re looking at?

care to elaborate on Julia? I was disappointed that a GC math-oriented language would have such a poor memory performance, but some stuff was nice, like the vectoring operations.

Reply on news.ycombinator.com