Use turbo mode (shift-click green flag) just in case. I mostly used events rather than custom blocks. It's not efficient since I only planned on making it as a quick project. I just started trying to run things on http://phosphorus.github.io/app.html?id=66064088&turbo=true&full-screen=true so try running it there. Click around on the screen to place some seed points that are not too close together; press When picking the number, pick 0 for the first 3 layers, 1 if you want a normal voronoi, and 2 or higher to see one that seems less directly related to the seed points. It will generate a lot of divided regions. The number must be at least 1 less than the number of seed points you have placed. After the screen is drawn, if you let go of space and press it again, the program will try to draw the next layer until it stops working..
This is really boring and poorly-written hard to understand rant so don't read it unless if you mind me making something similar other people's stuff without saying anything. I wrote the stuff below before I changed everything to be based off of changeable seed points and the project was still named "stacked voronoi." Inspired by Chladni figures and Voronoi diagrams. Maybe I'll also credit https://scratch.mit.edu/projects/20497702/ a bit just to be safe that no one is mad at me. Voronoi diagrams have regions consisting of points that are all closer to a certain seed (a few pre-picked points are called seeds) than any other based on which point places in the regions are the closest to. This program is different in that it distinguishes regions based on the top few seeds the points in a region are closest to, not just the closest one. The seeds for my project are all the points with whole-number x and y coordinates on the plane and only an eighth of a unit square is drawn, because the rest is just a repeating pattern. At first I was planning on making it based precise calculations rather than guestimating. https://docs.google.com/document/d/1kIGNc-XAgn7Qq4URghZCfSz0u1c6uxbzXZRcbMI-8SE/edit Also I looked at WO997's path detection
