[ODE] Object size, velocity, time step and collision detection.

[Adam D. Moss]

Сергей Зефиров wrote:
> Could ODE miss collision between two, say, spheres if timestep*their relative velocity is bigger
> than sum of their radiuses?

Yes.

> Is it worth of some effort to implement continuous collision detection, if so?

This topic has come up before...  there are various hacks to work
around 'missed' collisions, the easiest being to simply use a smaller
timestep.  While some cases of swept collisions are almost trivial
(e.g. sphere-sphere) in isolation, adding handling to ODE's existing
infrastructure might be quite ugly and inefficient without deep
thought -- even compared to the hacks.

--Adam
-- 
Adam D. Moss   . ,,^^   adam at gimp.org   http://www.foxbox.org/   co:3
"... so make sure your illness is early in the financial year."