🚀 go-pugleaf

>In article <Bx8H5G.Gt7@cs.vu.nl>, mgemmel@cs.vu.nl (Martin Gemmel) writes:
>> Writing a pinball simulator is what I've been thinking of for a long
>> time, too.
>> 
Could be fun.

>> f110019@hamlet.ucdavis.edu (Julian Burger) writes:
>> 
> } In other words, you would have a pinball-table description language which
> } would be parsed and a table created from it.
> 
A good idea, I think

>> The idea sounds good, but for the moment the most important question
>> is just how to calculate ball motions and the effects of the various
>> objects on the ball(s). 

The ball is moving on a sloping table, (slope in y direction). The x component
of the velocity remains constant. And the y component  vy(t) = vy(t-1) + a
where a is a fraction of gravitational acceleration. 
Two approaches to ball motion:
	i)  calculate the trajectories as you go, using the above equation.
		
	ii) precalculate the trajectories (I'm not sure how to do this, but..)
		Bresenham has a fast incremental algorithm for drawing 
		an ellipse that will step one pixel at a time. This may be
		faster than method (i) and guarantees a reasonably smooth
		trajectory. The ellipse would be a suitable approximation
		to the parabolic path of an object moving in a uniform
		gravitational field, I don't know what the best way to 
		initialise the algorithm would be, but I suspect a table 
		lookup would save calculation.
		Note that, although the shape of the trajectory is 
		pre-calculated, the actual position is calculated 
		incrementally, so there is no problem with collision
		detection. 
	
You could handle collisions, by treating each object as a collection of
surfaces of unit size, give each surface a position and an angle, and then
when the position of the ball coincides with that of an object then
calculate the deflected velocity of the ball.(I'm not sure about the 
trig formulae but I think it looks like:
	Vy(t+1) = Vy(t)cos(S + 180) + Vx(t)sin(S)
	Vx(t+1) = Vy(t)sin(S + 180) + Vx(t)cos(S)
). 
My guess is that only a few angles will suffice to give the appearance of 
infinite variation (particularly as the ball travels
in a curve), so just specify the angle as an integer multiple of the minimum
angle size (say 15 degrees) so the calculations can be done using a table
lookup for sin and cos. eg  VAR sin : ARRAY[0..23] OF INTEGER ; (scaled 
appropriately).


*		- Greg				E kore e ngaro, he kakano
*		gford@waikato.ac.nz		i ruia mai i Rangiatea