Hi everybody!
All of us, grasshopper users, played with kangaroo and circle packing at least once.
I’m sharing a little curious experience.
Probably all of this is totally obvious and trivial for experts in the field, and all my assumptions can be completely wrong.
(Sorry for my bad english 
)
Very nice explorations! Thank you for sharing.
I’d noticed the square arrangements showing up sometimes with oversized circle packings, but I’ve never really dug down into exactly why it happens.
It’s also interesting to compare this with points distributed through electrostatic type repulsion (see the examples here), where the repulsion strength is some negative power of distance. That produces large regions of triangular arrangements, but doesn’t seem to result in the other polygonal packings you get with the spring-like repulsion used in circle packing (where the strength is linear with distance).
I wonder if these square arrangements could even be used as a way of making quad dominant meshes. Perhaps there’s some way to encourage square packed regions to form even more.
Maybe a sort of multi-layered repulsion with 2 distances…
oh, cool!
I see what you did there… make sense…
yes that’s one of the first thing is suggests just by looking at it.
Also from your last picture: that’s definitively this
https://en.wikipedia.org/wiki/Grain_boundary
If i’m not mistaken, regular structure/crystals might tends towards some kind of advantages but also defects.
Like a too-much-uniform grid might be too brittle (NaCl) or very ductile (Fe)…
But, for example, by adding imperfections (carbon in steel) we can lose some ductility/plasticity to gain some hardness (and brittleness).
Probably all of this was already searched by FEM manufacturers…
When would we need a never-regular mesh? 
