> 1. There is SpheresContactGeometry. Is it supposed to be used _only > for contact_ interactions? > > 2. There is DistantSpheresContactGeometry. Is it supposed to be used > for _distant_ interactions? > No; the name is misleding; InteractingSphere2InteractingSphere4DistantSpheresContactGeometry creates SpheresContacatGeometry; contrary to IS2IS4SCG however, it will create SpheresContactGeometry even if penetrationDepth<=0.
[ I eventually want to replace IS2IS4DSCG by IS2IS4SCG with some flag that will permit distant interactions. There is no reason for such code duplication ] > > >if (penetrationDepth>0 || c->isReal) > > should be replaced by > > > if (penetrationDepth>0) > > as I suggested. And more, InteractionDetectionFactor should be removed > from SpheresContactGeometry. > Both conclusions are drawn from false premise, hence are not true. > > About Interaction's flags: I think, isReal and isNew should be left > unchanged, because their meaning is very clear. > isReal is not clear at all and I had to read sources (at all the places where it is used) to get to know what it means. isReal for distant contact is false, but the contact is not "unreal" nonetheless. (yes, isNew is OK) > > I propose to add a flag "cycle" and anticipate its use as a _local_ > variable in any Engine. I think such a flag would be useful in any > case. Its value it not supposed to be transfer from one Engine to > other or through iterations. I have used "cycle" instead of isReal for > mark stale contacts > Why not call it isStale then or something similar? Again, I suggest moving this discussion to the wiki page http://yade.wikia.com/wiki/Sanitizing_Contact_Logic so that we have some starting point and can decide what and how will change. Regards, V. _______________________________________________ yade-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/yade-dev
