FORTRAN floating point accuracy problems

[Pekis]

Hello,

Perhaps you should try these formulas:

For x>0:
DEG->DMS(x)=(90*x+INT(60*x)+100*INT(x))/250

It's mathematically exact, but special care is needed in evaluating INT(60*x) on a calculator.
=> It seems better to round 60*x to last internal digit before applying INT.

DMS->DEG(x)=(250*x-INT(100*x)-60*INT(x))/90

See you.