generative modeling for Rhino
I'm using kangaroo for some weaving things and i just find that it's really hard to make a rigid body in kangaroo. You should add some springs or Plaw.
When I find "Collidesurf", I just know that it will be a very useful tool!
But........................................the "Brep" that it support in Solid is really quite limited the pipe will cause a strange Bug that all points going to one position. Box is ok and I really want to know how about yours. All Brep is useful ?
I can't replicate this bug - can you post the file ?
Collidesurf should work with any Brep, so long as it is closed (it needs to be able to check whether points are inside or outside)
Ah, okay, I think I see part of the problem now.
CollideSurf is capable of either keeping points inside a volume (eg for packing problems), or outside (for solid collision like you have here).
The way it currently decides which one to use is based on where the particles start out at the beginning - if they are inside to begin with it keeps them inside, and if outside it keeps them outside.
When the radius of your pipes is going over a certain limit it is putting the points of that line inside them.
Would an option letting the user decide between containment or collision be helpful ?
So you could choose to make even points which start out inside stay on the outside (they would just immediately get forced to the outside)
Something else that might help will be putting in the option for force based collision. At the moment points get their position projected back onto the surface when they penetrate (and their velocity component normal to the surface reversed).
Force based would mean they just get pushed back towards the outside of the surface. This is slightly 'softer', and might help avoid the stability issues that seem to occur when changing Breps during simulation.
I am very interested in kangaroo's potential to create self-organizing systems of non-circular geometries. CollideSurf seems to have huge potential to create a system of packing where the packing geometries could be iterated through while maintaining the logic of the system. Creating stiff springs through an internal interconnectPoint of each component works well in some simple cases, but makes it tricky to do things like irregular sided polygons.
I'd love to review the pentagon packing file (http://vimeo.com/13853602) if you're able to share it.
Thanks for Testing
I know the inside and outside thing, but did you try to change the radius to 60(not changing during simulation, before that). All the points should be outside the Brep but they don't. So I'm not sure about the problem is from my computer or files. It just goes wrong when radius is 60(just 60!), whether above or below is OK and everything goes well.
I think an option is good, but we should keep that easy and don't affect the whole system. Because these tiny option may cause some paradox.
Hmm, very odd.
I don't think this issue is restricted to Kangaroo though - testing those points against those pipes for Brep inclusion gives false positives using the standard Grasshopper 'inside' component.