According to APR 4761A, safety can be defined as the condition in which the risk of harm to persons or damage to property is reduced to, and maintained at or below, an acceptable level through a continuing process of hazard identification and risk management.
This definition highlights the importance of identifying hazards and assessing risks in order to mitigate them and maintain an acceptable level of safety. It emphasizes the need for a continuous process of risk management to ensure that safety is maintained over time, as new hazards and risks may arise. Additionally, it recognizes that safety is ultimately about reducing the risk of harm to people and property.
IEEE Std-1228 (1994) defines software safety as "freedom from software hazard," where software hazard is defined as "a software condition that is a prerequisite to an accident," and an accident is defined as "an unplanned event or series of events that results in death, injury, illness, environmental damage, or damage to or loss of equipment or property". Here we assume that the term "property" also includes intellectual property.
Safety vs. Reliability
In many cases safety and reliability are in full accord. This happens when proper straightforward failure-free functioning of a system is enough for both reliable and safe operation. For example, an elevator's mechanical system: the more reliable the mechanical elements, the more reliable and safer the elevator.
Indeed, it is often true that the safer the system, the less reliable it is. Consider an elevator: The maximum level of safety is provided by an inoperative elevator—its doors won't shut on you or your dog; pressing buttons won't cause anything unsafe to happen. Enter the inoperative elevator, stay inside as long as you wish, exit it—you are 100% safe.
What about reliability? As the inoperative elevator is functionally ineffective, it's absolutely unreliable and unavailable for moving you up and down to different floors of the building— its reliability is zero.
To improve the safety of a reliable (moving) elevator, designers add elements and controls that limit and even decrease the probability of its adequate operation. For example, a sensor that indicates proper door closure may be added. If the sensor is out of order, the elevator won't move: reliability decreases while safety improves.
System Safety is a rational pursuit of acceptable hazard risk in which the system is treated as an integral part of a System-of-Systems , taking into account the interactions among system's constituent parts. System Safety is an integral part of the interdisciplinary approach of systems engineering and its pursuit of systems that meet stakeholder expectations.
The methods of System Safety are diverse and are driven by many factors, including:
System Safety Assessment relies on analytical results, in part due to the high cost of testing limiting the ability to rely on test-fail-fix strategies for designing a safe and reliable system.
Increasing system complexity, which makes it necessary to leverage both traditional and modern hazard evaluation mechanisms in order to identify and analyze comprehensively the full set of credible mishap scenarios that have the potential to lead to adverse consequences, considering all hazard causes and propagation pathways through the system.
The development of systems that operate at the edge of engineering capability, requiring a high degree of discipline in system realization and system operation management and oversight.
The use of unproven technology, requiring engineering conservatism to protect against unknown mishap risks while at the same time requiring allowances for novel solutions.
See also:
Safety Assessment
Safety Management
ALD Safety Services for Certification
SAE 4754A