The theory of best approximation and best error bounds is sketched. It is observed that spline functions provide the best approximating functions, and then an analysis is presented which obtains best error bounds for the use of two-variable spline functions in interpolation, differentiation, and integration problems.