error_tolerant_approach