Re: [julia-users] Re: Functions in degrees

[Steven G. Johnson]

The errors in sin and cos can be much larger than 1ulp for sin or cos of large 
phase angles (>> 360) in degrees, because pi is not exactly representable in fp 
while 180 is.