I had tried once before, in 2023, keeping the triangle mesh fixed, and moving the elevations. That didn’t work so well. I decided to try moving the mesh points this time. When mesh points get too close together, they merge, adding their elevations together. That creates spikes. So I then average out the adjacent elevations.
Part of this project required me to de-optimize mapgen4. In that project I made lots of assumptions about things that wouldn’t change … like the mesh. And the triangulation. And the simplex noise. And the order of vertices. But I wanted to change the mesh every tick, so I had to rip out a lot of those optimizations.
There are some bugs in this simulation, but I think the bigger problem is that things bunch up too much at the boundary point, and I want the mountain building to be spread out more. I gave myself a week to work on this project, and I will allocate another week at some point, but not right now.
Also, yes, there are a lot of diagonal line artifacts on the new continent. It’s something I noticed in mapgen4 as well, and at some point I will investigate the noise functions I’m using.
Also see diagrams on https://www.geologybites.com/mikesearle[1] for what I should be doing instead of what I’m doing here. I should keep track of the west plate moving underneath the east plate, and then pushing up the east plate.