[ODE] GIMPACT has been released with BSD license!!

[David Walters]

> Okay, one moment here. I really do appreciate this excellent
> contribution and all of your enthusiasm, but we can't just rip out
> OPCODE and drop in a new system.

Would it be possible to write a collider for GIMPACT tri-meshes vs.
OPCODE tri-meshes?

This would bridge the gap such that if GIMPACT was better for low-poly
dynamic tri-meshes and OPCODE was more efficient (at this time) for
the larger 'world sized' collision trimeshes that you would get the
best of both worlds...

Admittedly I would not claim the current OPCODE trimeshes are perfect
for static cases - sticky interpenetration of boxesat high velocities
against even flat surfaces are very easy to replicate and will
seemingly continue to be a problem until convex volumes are
implemented.

Anyway that is my idea.  Interoperability between GIMPACT and OPCODE
would seem to me to ease the transition...

Regards,
David

ps.Francisco you talked in an earlier e-mail about your initial
evaluation of OPCODE and some potential fixes you worked on relating
to winding order, etc. - are there any patches you can provide to
improve the current OPCODE in ODE?, it seems like you made some
improvements before you started GIMPACT and surely those can be of
immediate benefit?