/agdg/ - Amateur Game Development General

>>20368

Open Dynamics Engine.

It's powerful enough to be used in AAA games like Bloodrayne, and simple enough to power indie games like World of Goo.

Write your own 3D physics engine, jesus, this generation of programmers are so weak minded cuck millennials, you just want to use other people's work and claim as your own, JESUS.

I'm 25 years old btw

KILL

YOUR

SELF

FEGGET

FIGGET

FUGGET

MUGGET

REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

It's not hard to code physics, this is how you check for collision:

if (obj1.x + obj1.width > obj2.x + obj2.width && obj2.y + obj2.height < obj2.y + obj2.height)

{

// collision happens!

}

See? It's not hard

Learn

To

Physics

XZY

123

2 > 1

It's simple as that. I wonder what is the mental age of you people here.

Why do you even need physics these days. Most games don't even need that, and when it is needed, only really simple shit like kinematics. That shit's taught in 4th of ESO, for fuck's sake. Quake simulated forces half decently, and the only thing it did was calculate the distance to the center of the explosion and apply a velocity vector correspondent to it to the player controller. No special prefabricated whatever fancy shit.

Also >>20416 is mad in his head. Don't listen to him.

>>20416

>Write your own

>Provides the most naive implementation for physics collision

You're a moron.

To answer op's question, there isn't a good answer. ODE is smaller than Bullet, but getting there, Newton is probably the smallest, but like you said, it may not meet all your needs; but don't rule it out just because it's the simplest, if all you need is simple collisions, it might work for you.

I would be careful if I were you to not conflate codebase size with complexity, even though bullet is huge, it's really easy to use for most use cases.

>>20368

unigine

>>20416

>Yells op for not coding like hero's

>Use GML

>>20368

Bullet

>>20416

> weak minded cuck millennials

>I'm 25 years old btw

>you want to use other peoples work and claim it as your own

>that code

>FEGGET

>It's simple as that. I wonder what is the mental age of you people here.

>ironicshitposting.png

>>20582

The only other option that isn't ODE would probably be to roll your own

>>20368

also there is PAL, physics abstraction layer, it allows you to profile several libs and choose the right one

>>20417

>4th of Educación Secundaria Obligatoria

what the fuck is wrong with you

>>20416

Let say you mistyped and wanted to do that:

if (obj1.x + obj1.width > obj2.x + obj2.width && obj2.y + obj2.height < obj1.y + obj1.height)
{
// collision happens!
}

OK:

obj1.x = 5; obj1.width = 2;  obj2.y = 5; obj2.height = 2;
obj2.x = 0; obj2.width = 1; obj2.y = 0; obj2.height = 1;
if (obj1.x + obj1.width > obj2.x + obj2.width && obj2.y + obj2.height < obj1.y + obj1.height)
//Give: if(5 + 2 > 0 + 1 && 0 + 1 < 5 + 2)
//if(7 > 1 && 1 < 7)
// if(true && true)
// if(true)
{
// collision happens!
}

For the neophytes, a real good test would be :

//Works only if the edges of the cubes are parallel to the axes X, Y or Z
//If rotation, you are doomed
// And I do an if just for the sake to be readable.
// In fact you do directly a return !(what I put in the if) or return (application of De Morgan's law)
function collide(obj1, obj2) {
    if(
        (obj1.x + obj1.width  < obj2.x || obj1.x > obj2.x + obj2.width)  || 
        (obj1.y + obj1.length < obj2.y || obj1.y > obj2.y + obj2.length) ||
        (obj1.z + obj1.height < obj2.z || obj1.y > obj2.z + obj2.height)
    )
    {
        return false;  //No collision
    }
    return true;
}

By the way, De Morgan's law is:

not(a or  b) = not a and not b
not(a and b) = not a or  not b

>TL;DR

>Learn how the tests are done, read the algorithms

>But use a framework to do that to avoid stupid errors

>>20416

> 3D physics engine

> is actually 2D

> Does not support rotation

>>20624

>By the way, De Morgan's law is:

Derp. The same as simplifying any other polynomial, except its in binary (boolean).

Let Not be (* -1).

-1( a + b ) == -1a + -1b;

I swear, people in CS are so fucking easy to impress. They're also so full of hubris they take ancient communicative properties of mathematics and rebrand them as their own laws just because the numeric domain is base 2.

>>20636

If you're a math major, you should know that you can't expect properties to hold for every single mathematical object you use.

For instance:

xy = yx

That equation holds when x and y are numbers. But when they're matrices, it doesn't always work.

>>20636

>they take ancient communicative properties of mathematics and rebrand them as their own laws just because the numeric domain is base 2.

Fucking hell m8, de morgan was a mathematician working on creating propositional logic nearly 200 years ago along with Boole.

>Let Not be (* -1).

You seem to have missed that the and became an or and vice versa

If you want de morgans laws stated in math the following is how Halmos did it:

(A ∪ B)′ = A′ ∩ B′

(A ∩ B)′ = A′ ∪ B′

>communicative

kek, de morgans laws aren't simply restating commutative laws.

>implying propositional logic is base-2 arithmetic

Oy vey, read some set theory m8.

>>20656

inb4 >implying a universe, you fucking know what Halmos meant by complement

>>20416

>>20624

>>20628

>implying boxes

God is a sphere, get over yourselves.

I thought that that one god is a cube guy was a lone nutter, but I seem to have stumbled into some fucking box cult.

If you find your collision detection doing anything other than finding the distance between two points and comparing it to the sum of the radii then you dun fucked up and best repent as soon as you can lest you die in a state of sin.

You're perceived 'need' for boxes (and indeed, any non-spheres) is illusory and merely a product of your imperfect senses.

Being as round as I am, while far from spherical, I can perhaps aid you in realising your true need for a sphere.

Feel free to post your box-related problems here and I'll try to help you round them out a little.

Pic related is an animation with a framerate of \omega_{0} fps, showing perfectly without ambiguity a projection of a rotating Ω-sphere onto the plane. Behold, god.

corners are the devils work, do not be lulled by the edged and corners of your desire and sin, repent and accept the perfection of god-sphere

god.svg:

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!-- REPENT YOUR SQUARE WAYS SINNERS -->

<svg
   xmlns:dc="http://purl.org/dc/elements/1.1/"
   xmlns:cc="http://creativecommons.org/ns#"
   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
   xmlns:svg="http://www.w3.org/2000/svg"
   xmlns="http://www.w3.org/2000/svg"
   width="100%"
   height="100%"
   id="svg2"
   version="1.1">
     <circle id="god" cx="50%" cy="50%" r="40%" stroke="black" stroke-width="3" fill="black" />
</svg>

Playing with ODE and Newton now.

I might as well ask this as well;

Anyone know a good tutorial on 3D rotation physics?

I can do 2D sphere/square/tri but in 3D I'm lost.

>>20416

>just code your own 3d engine, pleb

>writes a 2d rectangle collision that doesn't even work

Please tell me this is b8.

>>20636

>Commutative

I think you mean distributive.

>>20656

>tfw latex is still broken

i'd say bullet. i'm currently interested in studying bullet 3.x and its fully GPU-based collision detection and rigid body dynamics.

if you want to read about it, i'd recommend checking out Multithreading for Visual Effects.

i'm also interested in studying it, and openVDB/opensubdiv, alongside some ptex. it's all really interesting stuff. i probably wouldn't be making a game, but i am interested in getting a nice engine out of the way first. at the very least i can also design the gameplay mechanics on the side while i work out the kinks.

on a side note, is there any FOSS implementation of something like YEBIS for post, and openAL for 3D proximity audio doppler stuff?

>>20368

Try this: http://urho3d.github.io/

>>22287

>>That's an entire game engine, not just a physics engine

>>Inspired by Ogre? Just fuck my shit up!

GTA V used Bullet, so there is your answer.

>>20368

Pure C++

Wish I would have discovered "Game Coding Complete" last year instead of discovering the caveats of game design through trial, error, and autistic optimization.

>>22387

As in you could code that yourself with just the <chrono>

<math> or glm libraries

you need vector representation, bounding boxes based off of it for collision detection, centers of mass, constraints, energy transfer equations

here's somewhere to start it seems

http://www.wildbunny.co.uk/blog/2011/04/06/physics-engines-for-dummies/>>22387

also there is this thread

>>>/agdg/17387

>>20416

This is seriously the BEST response in the entire thread.

Yeah the work is a little bit more, but it took me around 3 weeks to code my first engine and by the time it was done, I had already learned so much about programming, that doing simple shit like listening for input was menial.

>I got a job for a small studio after I put my game engine in my resume portfolia and now make 85k a year writing general backend code for systems.

>>22389

>but it took me around 3 weeks to code my first engine and by the time it was done

Show it to me, put it on github

>>22589

Bullet is super easy to use.