Just curious on the correct use of pi value. I have read at many places that tan inverse(1) * 4 would be most accurate, like

#define PI (4.0*atan(1.0))

or should I just copy and paste the well known pi value for 30 odd decimal places as a const long double directly from wiki.

One thing I know for sure is that the fraction 22/7 is not that accurate estimate of pi.