“…Thus the above problem can be reformulated as: Let f be a spherical convex function and let D=[x 1 , ..., x n ] be a set of discrete data sites located on 0 such that the data points f (x i ) x i , i=1, ..., n, are strictly convex, i.e., such that they are extreme points of their convex hull. Then there exists a C convex spherical function, interpolating f at D. In fact, as in Section 4.2, one can show that such interpolating functions can also approximate f. Finally, we note that some constructive methods for smooth convexity preserving interpolation can be found in [12,14].…”