Software reliability

Abstract

The paper discusses the basic concepts underlying the issue of software reliability, and argues that programmed exception handling is inappropriate for dealing with suspected software errors. Instead it is shown, using an example program, how exception handling can be combined with the recovery block structure. The result is to improve the effectiveness with which problems due to anticipated faulty input data, hardware components, etc., are dealt with, while continuing to provide means for recovering from unanticipated faults, including ones due to residual software design errors.