He's using a different sample set than I did, I only used the centre pixel points, but (I believe) he has also included points in between. Hence his version re

He's using a different sample set than I did, I only used the centre pixel points, but (I believe) he has also included points in between. Hence his version retains the blockiness of the Moore curve, whereas mine does not.