trekhleb / state-of-the-art-shitcode Goto Github PK

Always write comments in languages that are rarely used among team members. So that members can identify who wrote the code easily.

For example:

• Always use Elvish if you know how to
• If you're the only Chinese guy, then use Chinese
• If you don't know other languages, use family jargons if possible

chinese...

I know it's a ironical project, but I have question whether to write comments in native language.
Using Chinese will has trouble shifting input method, and it would lead to many problems between win and linux.
But using English will avoid these questions as well as practice Computer English, and it wouldn't cause any problems reading it.

Python and JavaScript allows it, so i think using English and changing layout hundreds of times is nonsense
Also some weird characters need struggling, so replace them via #define in c and do neater functions.

加个星，虽然我现在是个学生，但是以后可能会用到。

This project definitely violates the rule of not creating a README file.

Just a thought, what will happen when shitcode meets EE.

There's FizzBuzzEnterpriseEdition showing us how over-engineering and cargo cult would devastate your codebase in a hurmorous way.

It comes to me that shitcode is not only about how you write the code, but also how you choose the abstraction and implement it.

More "enterprise level shitcode" examples?

JavaScript allows for UTF8 variable names which allows use of some emoticons and some people even use emoji as identifiers in languages like C++ #39

Good : -

var ಠ_ಠ = 2022

Bad : -

var year = 2022

Nowadays many things support complex variable names. So why not name them in your native language? That saves your time thinking of a name, and time is important.

Good 👍

记录=提取记录(结果)

Bad 👎

processRecord = getRecord(Result)

Hey! I’ve having trouble getting GPT to work with this guide, any suggestions?

jk😂

他们说 代码写的跟shi一样 我特意来康康

I know that in some big company, dependencies don't be locked by lock file, because they need to apply hot-fix of upstream dependencies without having to update package.json manually.

Am I just blind, or is there no combined shitcode all in one file? I need this as an example.

Use iteration. not Loop. The best way use goto.
Must contain assembler，And comments say: It more fast. don't change!

I followed all the principles of this repository, and yet my game is very laggy and uses a lot of CPU and RAM. I don't think I can optimize my code further, but maybe one of you guys can help?

There should be an item regarding secrets in code:

Good practice :

• put your API keys and SSH keys in code. You will never forget them that way 😄

Bad practice:

• encrypt sensitive values or keep them in a secret vault

Don't indent, makes complex code take up more space in the editor.

Since there shouldn't be a README file, how can I use the badge?

The most important thing is that version control is not needed. Version control is time-consuming and unnecessary. It is recommended to delete the entire repository. 🤪

Looking at this README.zh-CN.md is like eating shit

chinese...

Based on WTF-PL. See

Good

String 你好 = "你好";
System.out.println(你好);

Bad

String hello = "你好";
System.out.println(hello);

chinese...

So is the auther Chinese?
是故意的还是不小心的？

Make your colleagues confused by your code!

Good:

a++ += ++a+a++;

Bad:

a++;
a += 2;