I have lost the ability to write shit code.

> starts new project
> it will be simple
> starts thinking about modularity
> starts thinking a out version control
> need to write an srs
> testing
> architecture???
> HLD and LLD
> can't stand shit code
it's over.

shit code is all I write
I just start fucking going and assume everything will work out and it does

>>106001370 (OP)
i got ya little bro

>>106001370 (OP)
That's analysis paralysis, write shit code already and deshit-fy it over time

Just tell yourself you'll fix it up later and then keep ignoring the fixing part forever.

>>106001370 (OP)
How about you write small scripts for your everyday usage

>>106001370 (OP)
adding
>// TODO: [some shit i will almost always never do, but when necessary i will not only do it but have context for why i thought i needed to do it in the future to start with]
whenever i start thinking about how complicated something might have to be in the finished product solved this issue for me. or in other words which i will now quote for you, "Only write what needs to be written."

As a side note, this also helps when you don't know what to work on next. Just ctrl + F "TODO"

>>106001370 (OP)
Then don't write the shitcode. Do all those management stuff and outsource the actual shit coding to the AI and call it vibe coding. It will be a well engineered piece of vibe coded shit.

Find a 'project based learning' tutorial: all those decisions have already been made, all you have left is write the shit code. Get practice. Get used to it.

>>106001370 (OP)
What do all these mean and how do I start learning them?

>>106001370 (OP)
just use AI. I am partial for grok.

>>106001370 (OP)
>> starts thinking about version control
you dont need good commit messages/commits until you have a basic minimally viable product to start expanding on
because there is nothing there the commits are all going to be massive additions and unstable modifications
there is no benefit to carefully documenting why you changed something because it was not stable or working in the first place

>>106001370 (OP)
>modularity
Fag
>version control
Straight line?
>srs
Shush
>testing
I'd rather shoot myself
>architecture
Who are you, fucking art vandelay?
>HLD and LLD
You need to take the DLD out of your ass
>can't stand shit code
Ironically you stand when you shit on the street, harjit

>>106001370 (OP)
>need to write an srs
I hate that. Just do modular out of the bat and it solves a lot of this.

https://www.youtube.com/watch?v=443UNeGrFoM

>>106005213
>how do I learn
I honestly started just asking AI to teach it to me and recommend me some external reading material or textbook so I can double check it's not hallucniating.

>>106001370 (OP)
Go back to roots.
Do something in assembly.

i have years of experience and i still struggle in first timing the design of a program. i generally have a right direction and just write all of it, not really caring if its the most optimal solution. then when im done i rewrite it so it will be maintainable. also how is version control an issue? i just copy everything every so often so if something goes wrong i dont have to deal with that shit. source files tend to be pretty small, so i can have hundreds of redundant copies and its no big deal.

>>106005213
Just look up lectures/books on (software) project management

>>106006624
>how is version control an issue?
You can't just shit out commits. You create branches based on what features your working on and if you're on your own, you have to switch branches just to fix shit so you can keep working on your feature and keep your workflow clean and traceable.

It's a pretty nice workflow if you want to abandon a feature half way through and still have working code.

>>106006996
Most of my projects had a linear roadmap. I start coding this, and can't code that until this is done and working. And when I finish coding everything, it works, I stop adding new features or shit. The feature list was decided before coding the first line. How does version control branching and stuff help me? Or is it because I only did toy projects and nothing large enough that required it?

>>106007049
Version control with proper branches just means you don't have to untangle a fuck tonne of code when you wake and realise your feature was a dumb idea.

It also helps you design code in was that ensures your features are decoupled.

I now do it for small projects. It was painful at the start and I was slow but I now accept it as fact.

I usually go down the merge and never delete a branch path rather than rebase.

>>106007070
Yeah, but my features aren't optional dumb ideas I can abandon. They are necessary stepping stones to reach the final product. If I drop one, the product doesn't meet the rewquirements. Most I can do is redesign the implementation, but if the code wasn't running it had no commit.

>>106007085
Anon you're building an inverted triangle. Feature branches don't have to mean things in the requirements doc. Just make a branch finish what you are doing then start the next one.

>>106007118
But if the branches don't actually "branch" out, they just go linearly one after the other, then why go thru the trouble of setting up branches in the first place? Isn't just shitting out commits more efficient?

>>106007085
I'm not shitting you or being intentionally contrarian. My setup works for me, but it seems like there is value in switching to yours, but I just can't see it.

>>106007142
you use branches because undoing shit is easier if things go wrong. even if you know you have to implement all of the fratures in the spec, sometimes you end up changing how you want to implement it midway through the implementation.

>>106001387
>crack is meth

small dick energy
always write jeet code then dejeetify it later