yade-users team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

> I would like to understand well the design you assigned to the
> collider. It is almost clear as you describe it in your documentation,
> just one more confirmation. If one does not include in the contact law
> to request the erase of the interaction for a certain condition, the
> interaction being real will always remain such. Is this right?
Yes. The collider is supposed to handle collisions (by creating/deleting
potential interactions). It is none of its business to operate on real
interactions.
> Is this design thought to facilitate the implementation of cohesive
> laws?
That was the original motivation. More generally, what law does is
completely orthogonal to what collider does. (One works only on
potential and the other one only on real interactions).
>  In my mind I was thinking I would assign to the collider the care of
> maintaining real an interaction or not. For laws that need
> adhesion/cohesion in the normal/shear direction we can still play in
> the contact law having enlarged bounding boxes. Would not be the same?
It used to be that way, and it doesn't work well. You need to have match
between collider and your contact law, otherwise it breaks down.

Enlarging all bounding boxes creates more potential interactions and
incurs (possibly severe) performance penalty. You have to adjust bbox
size to longest possible interaction.

If you feel like writing your own collider, that will only work with
your contact law, I would only say that it doesn't fit the way it is
currently done (which works reasonably well I think) and that it heads
towards the trend of everybody having his/her own half-assed classes.

What would be your motivation of not deleting contact in the contact
law?

v