--------------- EXPECTED PROGRESS ---------------

None - our schedule isn't drawn up yet.

---------------  ACTUAL PROGRESS  ---------------

- Added joystick support to dxInput

- Fixed many physics bugs ---
	- collision detection is now sort-of framerate-independant (see below for the caveat)
	- collisions are now checked and handled preemptively (as opposed to posthumously)
	  using a small step size, to prevent penetration (ahh, gutter-minded CS terms...)
	- collisions are no longer reported if objects are diverging (less stickies!)
	- impulse calculations are now correct and incorporate object mass and restitution coefficients

	Problem: I need to make sure that object movement is perceptibly independant of the
		framerate, while allowing for variable step sizes. I'm currently using the 
		midpoint (or RK2) method for integration, which is a very poor approximation --
			 
		1) if I make the step size fixed, then movement will be framerate independant. however,
		on really fast computers, the render rate may exceed 1/step, which would effectively 
		limit the visible rate of animation and places unfair performance constraints on the 
		physics code (BAD!)
		2) if i don't use a fixed step size, and instead just put an upper bound on it, then 
		there's no jitter, but on fast computers objects will move at slightly different 
		speeds (which will either cause a much bigger kind of jitter as the network syncs, 
		or give fast computers a slight gameplay advantage - also BAD)
			
---------------       SNAGS       ---------------

No schedule snags...

---------------       PLANS       ---------------

- implement networking!!!
- write simple action mapping module for input
- figure out how to implement runge kutta order 4 integration, to fix the step size issue
- finish and test the polygon collision algorithm:
	a) bounding poly has to rotate along with the ship
	b) must calculate poly/poly collision normal