An analysis of the famous catastrophe Therac-25 pitfalls, a computer-based radiotherapy device developed by AECL (Atomic Energy of Canada Limited) between 1985 and 1987.
This analysis is based on Mrs. Leveson's investigations article posted on the Stanford web archive with some Hardware-Software Co-design in mind.
- What's Therac-6, Therac-20, and Therac-25?
- The "Therac" series software defects.
- The "Therac" design pitfalls.
- How hardware protection overrides software pitfalls in Therac-6 and Therac-20.
- Therac-25 reported
Malfunction-45. - A deeper look into a simulation code and simulation fix environment.
- Naming anti-patterns utilized in the Therac software (Race conditions and concurrent programming).
- An approach to avoid these catastrophic errors.