Floating point calculation programming principle, implementation and application source code