A method to determine test set thoroughness by measuring the extent to which a test set can discriminate the program from slight variants of the program.