On this website (http://www.ualberta.ca/~pogosyan/teaching/PHYS_130/FALL_2010/lectures/lect35/lecture35.html) it says that the formula $\sin(\theta_{max}) = (m+\frac{1}{2})\lambda$ derived for the maxima of a single slit interference is only approximate, while the formula for the minima $\sin(\theta_{max}) = m\lambda$ is exact (if we neglect other approximations like interference of parallel light waves). Why are the maxima not just in the middle of two minima?
We can clearly divide a light ray in three parts from which two cancel and thus get the mentioned formula for the maxima. Why is this formula not exact?