Physics and Programming

Implementing physics in programs is fuck easy.  Let me illustrate:

Each object has an x, y, and z coordinate.  Per frame it exists in a particular x,y,z coordinate.  To model forces acting on that object, we need two attributes for each coordinate: current velocity (v), and change in velocity per frame aka acceleration (a).

For each frame, using x as an example, do this:
 x.v = x.v + x.a
 x = x + x.v

When forces act on an object, you have to modify a and v accordingly.  Velocites are reduced by division, and directions are reversed by a sign change.  So, say object A and B collide, you can model that like this:
 
 A.[coordinate].v = - (A.[coordinate].v / 1.5)
 and same for B
 
You might want to divide by a factor dependent on the objects weight or mass, so you could define that as well, a w or whatever.

Gravity is easy too.  Just do something like this per frame
 [object].y.v = [object].y.v + .05
 for .05 substitue a weight factored value or other value that works depending on the types of objects

IT'S THAT EASY.

------------------------
Can someone elaborate more on this. Or give more examples.

This is all rather naive. Don't do it per frame, do it per second, so you have smooth animation if the framerate jumps around. For example:
Gravity
obj.y.v += sec_diff * 9.80665
where sec_diff is the time-delta since the last frame, in seconds.

>>2
Teach me more please.

>>3
Change all the other formulae to similar things involving sec_diff.
Except the collision one, scrap that entirely and take an AS-level physics course. Or just use one of the countless thousands of collision models out there.