Part 4/6:
In the 1980s, there was a prevalent belief that once software was functional, it could not fail. This mindset led AECL to allow a single programmer to code the Therac-25's software in PDP-11 assembly language, with little oversight or rigorous testing.
The software's design was unnecessarily complex, and when questioned about the testing process, AECL representatives provided misleading information, claiming the Therac-25 had undergone 2,700 hours of testing, when in reality, it had only been tested by operators using the machine.