The sum() example does not work as the article says it does. Under Visual Studio 2013 compiling for x86 (C++ compiler, but C++ also has integer promotion rules), the function returns zero. The reason is: 65535 = 2^16-1. uint16_t is signed, therefore it has 15 bits to represent the value. When executing "int16_t a = 65535;" under a debugger, "a" is set to -1.