Something I learnt over the years is that it doesn't matter how fast your PC or Server is, people will produce a piece of software that will be capable of running slow on it.

Not weird with languages getting worse: https://youtu.be/U4c6nFGt1iM/...

Only Rust of "new" languages performs like the old ones. But fuck Rust

Yes, yes I can, and will.

@retoor I have a problem with saying Python is soooo slow. If you know how to program it properly and push most of the work to the internal side or libraries that are compiled it will do most of the work there. A lot of the time the libraries are compiled C/C++. The problem is most idiots program Python like they did in C. The strength of Python is gluing all these libraries together.

@Demolishun I completely agree. If you use the language correctly, and you don’t try to recreate the wheel, then python really isn’t that slow

@Demolishun I agree, also, while writing my own language I found out how fast python is. It's hard to match the speed. Since languages are made for humans and not computers, I consider python in syntax maybe the best because it's the easiest. But if I could smash a button to move the whole world back to C, I would smash it 😂

@Demolishun anyway, that video is good. Java on #4. Who expected that

@retoor I think there is a subset of Python you can write and compile it somehow. I can't remember what it is called. Python modules are also not hard to write. You can always isolate the heavy code and push it to C.

Ironically the coolest library I have used for Python is called Construct. It is pure Python, but very useful for building and deconstructing streams of data. A cool feature is once you build the stream and you can invert the operation and deconstruct the stream using the same object. They used to have a text portion in the library and it was what I actually used to process some serial stream data. It was a lot of fun to work with.

Doh, link: https://construct.readthedocs.io/en...

@retoor Honestly Java is an amazing language due to the technology in the vm. The coolness factor is high.

@Demolishun cool stuff. you can always compile to pyc. Don't know if you mean that.

I did write a python module in C for fun. The extensibility is amazing. But Lua is even bit better regarding that. (For implementation in your project, writing modules for, I dunno)

My favorite python lib is dataset: https://dataset.readthedocs.io/en/...

Database as dict with the built in sqlite as backend. Database structure on the fly

@retoor no, there are ways to compile to real code using libraries and compilers outside the interpreter. There was one library that let you use decorators to specify functions to be compiled. You couldn't use all the features of Python for those functions.

Oh that is a nice library.

@Demolishun the numba @jit decorator probably. I tried it, it only supports basic python (and only native types?) indeed

software gets slower faster than hardware gets faster