Compiler Selection for Functional Safety in Robots

If a system developer can demonstrate that as much of its application source code base as possible performs as intended during requirements-based testing, then the likelihood of field failures is low. While this statement is largely true, due to the focus on high-level source code in many applications, such a strategy relies on the compiler’s ability to generate object code that replicates the exact intentions of the developer. But what if the compiler is flawed?