Introduction • point rendering • how adapt point densities? • for a given viewing position, how can we get n points that suffice for that viewer? • how render the points? • given n points, how can we render an image from them?