1) Why do we need 4x4 matrix for perspective projection ?
2) Why does a clipping happen in homogenous coordinates and not in NDC (Normalized Device Coordinates) ? What is the algorithm for polygon clipping with homogenous coordinates ?
3) What does it mean to do a perspective divide ?
4) During rasterization, for perspective correct interpolation to happen, Is it the inverse of NDC z-coord that gets interpolated or Is it the inverse of the camera space z-coord ? Since rasterization happens after the perspective divide at which point the camera space z-coord is lost.