Re: [ODE] Russ: [..]making the iterative solver fully stable[...]

[Achim Moller]

Re: "so that the solver can be warm started."

I have read somewhere that warmstarting does not help very much if you have
a fast moving "constraint situation". Getting tons of _different_ contact
points per frame is not very suitable for warmstarting, right? My
experiments (with the warmstarting code in ODE) did not show any improved
stability here.

Any comments?

Thanks,
Achim