[ODE] Contact Constraints: First Order Slip?
russ at q12.org
Thu Nov 28 13:43:01 2002
> What about finding a way to generalize material properties such that
> contacts could be generated automatically?
the trouble is, there are many schemes for generating contact parameters
from material properties, and most of the useful schemes are rather
ad-hoc. it's impossible to come up with a single scheme that will please
everybody. the near-callback provides a convenient place to compute
contact parameters, and this is also where contact culling decisions can
the alternative to a near-callback is a "get list of all potentially
colliding objects" function - but then you have the problem of the
caller having to allocate a maximum size return array. consider the case
where only 10% of near object-object cases will lead to contacts - the
user is required to allocate an array of 10 times the potential maximum
number of object-object intersections - not always an easy thing to do.