[ODE] Rope Simulation

[Pål- Arve Nilsen]

> Lower the suspension point of the rope (the attachment point) without 
> adding another segment, until you have space to insert another 
> segment; then you attach the segment to the top of the rope and the 
> segment to the top of the ceiling again.
>
> Note that you don't need to render the rope the same way you simulate 
> it; you always render it going all the way up.
This is what i have done for my rope and it works really good. Thoo I 
have big problems with making the rope stable. It tends to explode when 
it's streched, the ball joint gets pulled apart.
I actually found the fixed joint to be most stable. I have implemented 
my own version of the fixed joint, where i have added the possibility to 
set the CFM and ERP individually for linear and angular motion in each 
axes. This way i can simulate ropes with different strecthing, bending 
and twisting properties. I have limited testing with this joint, but 
from what i have testet i have gotten the best result from this joint.
The big setback is that too many joints really kills the performance 
since im actually setting 0 dof between the capsules jointed together.