It makes sense to use a degree-2 parabolic curve to connect the point and the neighboring 2 points. When you take the derivative, you get a linear function. Hence the (p_3 - p_1)

It makes sense to use a degree-2 parabolic curve to connect the point and the neighboring 2 points. When you take the derivative, you get a linear function. Hence the (p_3 - p_1)