I can’t believe May is almost gone. Still feels like so much more needs to be done. That’s probably because there’s so much more to be done but hey, I’m getting there. : )
I’m making sure I live up to the hype.
In regard to last update I was talking about how deep I would go with physics, after a few days of work I can say I can go a lot deeper than I thought. Thanks to one “olii” at gamedev.net I’ve managed to see that the separating axis theorem is much more powerful than I had first thought. Thusly, the collision can get a bit more wily and be more stable than ever. The physics aren’t in place to manipulate objects more realistically but instead to add more dynamic objects and platforms and such without error. Ohh the possibilities of objects in a dynamic world!
Also, if you have sphere-to-triangles collision working, you can do alot with spheres!
For example, a snowball, rock, egg, huge pine cone, you name it, can be pushed around linearly over triangles/ground. Plus a rotation matrix can be made using that pushed distance and the character side-vector, to apply “roll” to the rock etc.
And most characters can stand on a rock, or snowball, like a dynamic platform, and stuff.
Man, for me that SAT theorem was a bit hard, I read about it before but decided to code an intuitive but slower function. The people at GameDev are really nice though.
Actually, I was always using sphere to triangles collision, just using a different method. Ellipses to be exact.
Every moving object in the game from the characters to the boxes, to the 2D weapons was made up of ellipses. It was a very limited setup.
GameDev is a nice place : )! Honestly without GameDev I’m not sure where I’d be.
The ellipse method seemed pretty cool. So cool I actually tried implementing it now and I think I got it right – getting collision normal and depth back from transformed “space” was pretty difficult.
I had an OBB around the player character, before, but that wasn’t as smooth.