Common Math Library
You know what's weird. We've been making games in c for like half a century.. and there's still no consensus on Vector2, Vector3, Quaternion, Angles. Everyone re-makes these things for their own engine, or their own middleware.
Then when you use a middleware library for physics, or audio etc, you end up converting between their Vector3 and your Vector3. Loading a model using a middleware library? Here you go, here's a new Vector3 implementation. Using a third-party navmesh generator? Here you go, another Vector3 implementation.
Why is this like this? Why are we all speaking different languages?
I get it. We can't trust the c++ standard library guys to implement ANYTHING without it being an ordeal with everything over-complicated, templated to fuck. So obviously we shouldn't trust them.
But we have the internet now; we have GitHub, we can self-organise. Why hasn't someone said - hey we're making a common math library, we're going to keep it simple, we're going to accept code, we're going to do all the simd stuff that you don't understand, it's gonna be mit licensed we want everyone to use it. Maybe it's even all header-based; people love that.
Are we going to keep writing the same 100 functions over and over again, forever?
Then when you use a middleware library for physics, or audio etc, you end up converting between their Vector3 and your Vector3. Loading a model using a middleware library? Here you go, here's a new Vector3 implementation. Using a third-party navmesh generator? Here you go, another Vector3 implementation.
Why is this like this? Why are we all speaking different languages?
I get it. We can't trust the c++ standard library guys to implement ANYTHING without it being an ordeal with everything over-complicated, templated to fuck. So obviously we shouldn't trust them.
But we have the internet now; we have GitHub, we can self-organise. Why hasn't someone said - hey we're making a common math library, we're going to keep it simple, we're going to accept code, we're going to do all the simd stuff that you don't understand, it's gonna be mit licensed we want everyone to use it. Maybe it's even all header-based; people love that.
Are we going to keep writing the same 100 functions over and over again, forever?
https://xkcd.com/927/
I wondered about missing normalization and standardization for months when i started to learn programming. I asked myself every fucking time why theres no universal truth, but why every damn GIFTED Human that can programm complex stuff that the average joe can't has to be so completely out of bounds with his implementation like hes good at logic but lacks common sense.
After some years i just can guess its because they can't think forward. Theres people who live in the NOW and not in the future. They think if they cook their own soup, use their own definition of variables, math etc. it would not matter in the long run.
The problem is you can't tell them. Those people are often the most detached from reality autists where you can be happy if they release ANY documentation on their work let alone explained in a understandable manner. And the worst part is if you call out something absolutely logical and rational the people around them will defend that persons work like "But it works"... yeah it works but its completely unnatural, uneeded hard to learn and needs translation all the time from the general approach.
The biggest problem is imho no matter how much "working junk" someone produces, if you REALLY have a better approach and would want to add a change or improvement, nobody will listen to you.
Add a Comment