This paper describes an approach to generating reference data sets to evaluate the performance of algorithms used in coordinate metrology for form and geometric tolerance assessment. The approach starts with the reference results, e.g., the solution feature parameter values, and then determines the coordinate data. In this way, the expensive development of reference software is avoided. In this paper we consider the generation of reference data associated with determining the best-fit surface according to the least squares (Gaussian) and Chebyshev criteria.