From the documentation of sind
:
For integers
n
,sind(n*180)
is exactly zero, whereassin(n*pi)
reflects the accuracy of the floating point value ofpi
.
So, if you are extremely troubled with the fact that sin( pi )
is not precisly zero, go ahead and use sind
, but in practice it is just a wrap-around sin
so you actually add a tini-tiny bit of overhead.
Personally, I prefer the elegance of radians and use sin
.