How to use Codon, a new alternative to Python

preview_player
Показать описание
The Codon project offers a highly Python-compatible language, built from the ground up, that compiles to native machine code for speed. It's not a drop-in replacement for Python, but it promises some significant speedups for math-and-stats applications. Could Codon's approach be how Python gets faster in general? Or is Codon more of a complement to the existing Python ecosystem rather than a long-term replacement? Watch and learn more.

------------------------------­----
Рекомендации по теме
Комментарии
Автор

A very interesting theme for next videos would be the performance benchmarking of the different solutions that attempt to compile Python into binary code.

HiltonFernandes
Автор

Just learn Nim and program in it if you want to be almost as productive as in Python.
Learning curve is orders of magnitude lower as compared to learning Rust and it's also way more mature than Zig.
You get C like speed, productivity close to Python and metaprogramming superpowers close to what you can get in Lisp, you can choose multiple memory management approaches depending on your goal and constraints and you also have way better interoperability with C and c++ than by using python while also being able to leverage python libraries if you really need them.
And once work is finished on incremental compilation a language like Julia becomes kind of obsolete.
Nim already has fast compile times but once incremental compilation is done you will get crazy small buildtimes, similar to how a interpreted language behaves or jited language like Julia(in addition multiple dispatch is such a big thing in Julia and I've heard multiple language designers saying that it is a bad idea to have that kind of unreliable mechanism at the core of the language ).

encapsulatio
Автор

The solution is very easy, just let AI like GPT4 annotate, type hint automatically the whole software stack, all of the libraries as well. I am not sure if it is possible to do this with very high accuracy, but I was testing it and it never made a single mistake, so maybe it is not that crazy, especially with upcoming AIs that will get even better.

Aldraz
Автор

Yes. We need to head in this direction. We shouldn't need to manage libraries after the script is complete.
There's a reason why containerization is taking off.

eazolan
Автор

I learned how to use one at school. Some educators came in and taught all of use. We even practised using bananas.

JeffBourke
Автор

Big pep to hear Codon delivers native binary better safety from reverse engineering.
1) Will pycharm support ?
2) Intentation is much like python ? (which has no logic to follow strict rules in intentations.)
3) Can we import current complex python code on Codon directly and start using ?

StartNight-dfsv
Автор

I love these concepts but to me they'll just be a fringe until I can import anything I want from my anaconda(or some codon equivalent) distribution and not have to worry about the code breaking or running slow.

glenneric
Автор

I can't understand the installation tutorial on github. . .😔

rufus
Автор

Great explanation! Can Codon work with FASTAPI? If so, what would be required to use Codon in FASTAPI, instead of Python? And, is Codon compatible with Pydantic?

MrAverageViewer
Автор

Thanks, do you know if Codon will be ever available in Windows?

alirezarasekh
join shbcf.ru