[ODE] Brakes and rotational damping

[Deak Szabolcs]

Matthew Harmon wrote:

>I guess my real question boils down to this:  Is there anything
>fundamentally "wrong" with taking the angular velocity of a body, reducing
>it, and putting it back?  It seems like a kludge, but that sort of how
>brakes work in the real world.  They don't apply a "counter-torque" as such,
>they bleed off momentum into heat.
>
Hi Matthew, 

(and hi list btw, I believe this is my first post though I've been 
around for a while)

Although Jon & Chris have mostly answered the problem I'd like to add a 
comment specifically to this question.
Yes, there is a problem with simply reducing angular velocity, because 
brakes *do* produce "counter-torque". (I left out the effect in my first 
attemp when switching to ODE and my car sim handled like Colin Mcrae 
Rally ;) )
Consider this: the angular momentum of a closed physical system cannot 
be changed by purely internal forces (nor the linear, for that matter). 
If two "floating" bodies are attached by a hinge joint one of them 
cannot apply torque (accelerate or decelerate) to the other without 
itself being affected by a counter-torque. This is actually what 
dHingeAddTorque (if I recall the function name well) does: apply the 
same torque to the two connected bodies with opposing signs.
Forgive me if this was obvious to you and I misunderstood your mail :)

Szabolcs