yade-dev team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

> Rename as you like. I only care about the equations.
I don't really have preference here, I was asking what sounds more
intuitive.

> One remark for consistency of the names : even if we find a good name 
> for SomePhys (FrictionCont? ElasticFrictionCont?), we still don't have a 
> unique way to convert FrictionMat into FrictionCont, and it should be 
> reflected in the functor's name perhaps.
> Actually, I've put the word "Simple" here because there is no complex 
> micro-macro passage in SimpleElasticRelationShips (just kn=E/D, 
> ks=poisson*kn...), as opposed to the older 
> MacroMicroElasticRelationShip, which was the law inherited from 
> Frederic, Hentz, et al.
> SimpleElasticR is not _the_ good way to convert FrictionMat to 
> FrictionCont, it is just one way, and at one point somebody will want a 
> different conversion (Frederic's one, Hertz-Mindlin,... they all go from 
> FrictionMat to FrictionCont).

Yes, that's very true. We can either have
Ip2_2xFrictionMat_FrictionPhys_{different things here}, or have just one
class with an attribute specifying which formulas to use.

>  "Ip2" sounds like other types of 
> assumptions are excluded. Or perhaps you see a different approach in 
> this situtation?
Ip2 just says it is 2-argument functor creating an instance of
InteractionPhysics.

v