"Rasterization = sampling a 2D binary function" This is a great way to explain rasterization. When I learned the definition for the first time, I learned it merely as the process of filling the triangles. Sampling and reconstructing(encoding and decoding) put so much more context into this process in my opinion.

With this definition: "Rasterization = sampling a 2D binary function", does it mean that ray-tracing is also one way of rasterization?

Essentially the f(x, y) here is a ray from the camera towards the scene, and if the ray hits any triangles, our inside(tri, x, y) will return True. And to calculate the color, we could bounce the ray multiple times after that, but eventually it will return one single value for image[x][y]

@ethanyanjiali re; RT, I think so. Just that in RT we'll update the value of the pixels every time there is a collision. And, it will be recursive with some kind of decay to ensure convergence.