Can we certify software that prevents life-threatening accidents?
In short: Modern aircraft and cars rely on millions of lines of code and dozens of processors. As software increasingly replaces human operators in critical roles, the potential for software failure to become a leading cause of accidents grows. But are current regulatory standards actually based on sound computer science?
In safety-critical systems, the stakes of a failure are significantly higher than in safety-related systems. While a failure in a domestic oven's temperature controller is a safety-related issue due to relatively small risk, a failure in railway signaling and interlocking is safety-critical because it carries a much higher risk of injury or death. Professor Martyn Thomas explores whether the standards used by regulators to certify these systems align with recommendations from the US National Academies regarding certifiably dependable software. This lecture examines how these regulatory frameworks compare to the US National Academies' guidelines for dependable software. It specifically investigates if current certification methods are grounded in robust computer science principles.
The complexity of these systems is compounded by the fact that accidents rarely stem from a single isolated cause. Instead, they often result from a combination of multiple contributory factors. For instance, a car accident might involve a combination of speed, road conditions like oil patches, and the placement of road signs, all of which may be linked to a chain of events involving maintenance, design faults, or human error. Consider a motorist skidding on oil while rushing to a meeting. This scenario illustrates how various elements, such as road sign placement or poor maintenance, can converge to create a disaster.
As programmable systems become more prevalent, the scope for software-driven failures to act as a leading contributor to accidents increases. This shift challenges engineers to apply established safety principles to software-based systems, a task complicated by the difficulty of tracing a single cause through a web of interconnected human and mechanical factors. Engineers must navigate these complexities as software takes over roles previously held by humans. This evolution makes software reliability a vital component of modern safety engineering.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Summary by Intellect · source: Safety-Critical Systems - Professor Martyn Thomas CBE
- Also: Safety-Critical Systems
To cite this card, link to https://intellectplus.co.uk/app/ideas-safety-critical-systems-professor-martyn-thomas-cbe.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
