geometric problem

[anubis]

hm, i can't get my head around this.
i have cube sitting at the origin of the world. i want the cube to emitt particles on it's edge points against it's rotation and movement. my idea was to safe the modelview matrix of the object compute the new one and use them both to get the old and new position of any given vertex emitting a particle. the direction in which particles get emitted would be the vector from the new pos to the old one. somehow this doesn't work though. is there something fundamentaly wrong with my assumptions ??? does anybody have a better idea ???

[Dia Kharrat]

if you are dealing with only position, why would you need a matrix? Why not have variables that store the old position and new one?

[anubis]

well for more complex objects than a cube it would be waste of memory to store that object twice

[baldurk]

btw (OT) apex: you should put your avatar on devmaster.net

[fringe]

hmmm, makes sence to me you basically do this right?

Normalise(matrixBefore*corner - MatrixAfter*corner) is the direction of the new particle


fringe

[anubis]

and that's exactly my code... something is wrong... i'll check the math code. maybe the bug is there

[fringe]

I can have a look at the code if you liek but definitely no promises. Send it to,

holtjr *at* th.ph.bham.ac.uk (<----trying not to get emailed by bots)

fringe