I don't quite understand the discretisation error of Euler's method
$$L(t,h)=\frac{x(t+h)-x(t)}{h}-f(t,x(t))$$
What I don't understand is: $\frac{x(t+h)-x(t)}{h}$ is the "gradient" and $f(t,x(t)$ is the value of the function. How can we get the error by subtracting the former by the latter?
Thanks.