ok i shouldn''t say "stupid" as hermite interpolation is useful but not for what i''m doing. ok i''ve been doing a little experimentation with the hermite interpolation and vectors in DX. well i''ve run into a bit of a problem. i want to make a sort of camera scripting thing, where the path is defined by nodes, and then the camera''s path is smoothly interpolated on a continuous arc through these nodes. using hermite interpolation, and setting the tangent vector for each node to be the next node in the sequence, i got it to look very close to what i wanted. in fact it moves in the exact pattern i want it to move, but not the right speed. you see, because it uses a parabola to do the interpolation, it slows down near the nodes and speeds up when it''s between them. so it''s a very obvious speeding up and slowing down as the camera moves from node to node. i was wondering if there was another type of interpolation that would move in an arc, like hermite, but that would not slow down near the endpoints of the arc! __________________________________________ you just think i''m here. i''m really not.

scale your velocity along the u axis by the reciprocal of the unit arclength at the current point.

My, that made no sense. Lessee. Formula for arclength is:
Integral r to s of (sqrt((h''(t))^2 + ((g''(t))^2)dt)
So the "sparseness" of the curve is:
s(t) = sqrt((h''(t))^2 + ((g''(t))^2)

So each frame, you have a current t(n-1). So your new t(n) is:
t(n) = t(n-1) * (1 + v/s(t(n)))

This formula is an approximation, but if you move slowly over t, you should be fine. You''re on your own figuring out the derivatives; it''s been a while since I was in calculus.

And please, someone call me on it if I''m talking out of my ass.

But... but that''s what HITLER would say!!

i think i might know what you''re talking about kinda i''m not in calc yet (hell i''m not even in trig yet) tho.

as a sort of compensation for the slowdown, i made a little loop that kept moving the object along the arc until it had moved a minimum distance (like 1 unit). unfortunately this didn''t work very well; in fact, depending on the minimum distance i had it set for, either the problem got WORSE or instead of slowing down at the nodes, it sped up!

maybe there''s a formula out there for making "linear" arcs..?