THE TYPE OF TESTING THAT CHECKS FOR ERRORS AT THE BOUNDARIES OF PROBLEMS