Few points that I've not seen being factored in over the past decade-

* This needs a better renderer in today's day & age

* Need cross-device/web support

* Topology Optimization w/ pure physics code

-----

Hopefully LLMs can work on forking this or adding better features with AI-assists