# Question in measurement uncertainty calculation in projectile motion

For a projectile motion, $$\displaystyle t=\sqrt{\frac{2y_0}{g}}$$.

If you look at page 6 of the article(http://www.hep.vanderbilt.edu/~maguirc/Physics231/p231_lect3.pdf), the fractional uncertainty/relative uncertainty was calculated by $$\displaystyle \frac{\sigma_t}{t}=\frac{1}{2}(\frac{\sigma_{y_0}}{y_0})$$.

However, at page 8 of the article, the usual calculus to calculate absolute uncertainty: $$\displaystyle \sigma_t=\frac{1}{2}(\frac{\sigma_{y_0}}{y_0g})$$.

These two results does not agree with eath other, (Notice: $$\displaystyle \sigma_{relative}=\frac{\sigma_{absolute}}{value}$$).

What went wrong? which was the correct form of uncertainty? and why?