Hi Grasshopper community!
I’ve been trying to write a script that generates geometry similar to the diagram below. The concept is that there are positive and negative “ions” that determine how the “rubber band” then loops around them. I am a little bit stumped off the bat but I have managed to get the circle tangents to connect but this is quite far from what I am looking for…
Any suggestions or ideas on how to approach this challenge would be greatly appreciated 
Line.TryCreateBetweenCurves method (rhino3d.com)
This is a good starting point to get the tangent lines between the “ions”. And here is a discussion on a similar topic: How to find external tangents between two closed planar curves (not circles)? - Grasshopper - McNeel Forum
The positive/negative sequence cannot be arbitrary, it mostly is a consequence of the angle of the corners of the starting polyline (passing through circle centers). Then, if the circle radius is big enough it can “flip” the shape in some places… but mostly you could pre-determinate the positive/negative pattern by just measuring the angles…
Kangaroo? Don’t have the time to try right now. maybe later. 
We have not the same logic, yours is very good. For the moment my little component gives that.
That make sense.
It’s a totally different scenario:
…yours do a task more!
Thanks for the help @maje90 & @laurent_delrieu!
This is fantastic 
Hello
I added some components in Nautilus plugin. There is still my old method and the one from @maje90 . As I wanted some automation to have some quite good “polyline” I implemented a pseudo TSP from a Mesh. I say pseudo because my tool doesn’t take all the points of the mesh. But it is quite cool to generate rubber band from points.
The tool is in Nautilus 0.9.5 version that is on the Package Manager
RubberBandAndTSP.gh (11.1 KB)
That is great Laurent.
What if the turning points are ellipses instead of circles?
standard sizes? ellipse orientation?
I want to know if a “rubber-band” wrapped around both changes length when I rotate them? (blue group)
P.S. Actually, I’m wondering if I rotate one of them ± 45 degrees, what is the effect on the other one, if driven by the rubber-band? I believe I can figure that out without the rubber band but if it changes length, the idea is doomed. Thinking about alternatives to this Ackermann geometry for catamaran rudders where the “outside” rudder turns less than the inside rudder.
test with convex hull and the length does change from ∼84 to ∼85
Aha, so it does. Thanks, good simple test.
Whoa, how did I get in the “Grasshopper Developer” category? I don’t belong her. 
