Been a while since I posted something here, I was too swamped with work stuff the last months. Anyway, I had this idea for a DLA-based particle simulation to create snowflakes tonight and here are some results. Some of those don’t resemble snowflakes at all, but they still look kinda pretty
The algorithm I implemented is quite similar to the one I explained here: Exploring Diffusion Limited Aggregation in Geometry Nodes
The only difference is that it is constrained to 2D space, and each iteration mirrors and arrays the newly added points.
As well as symmetry/tessellation, have you considered trying to incorporate elements of self-similarity, for example, each successive iteration also includes a miniaturized version of the previous one? Snowflakes are famous natural examples of fractals, although the exponential increase in points might harm performance without other considerations.
Thats a very interesting idea! I think the setup should be a different one in that case, but it could yield interesting patterns regardless. It might not be possible to “grow” these as easily, since each frame would add another layer to the fractal. Could be solved with a repeat zone instead. I’ll have to think about this
