I have read that the derivative of a function $f$ at a point $a$ is given by $f'(a) = \lim_{h \to 0}((f(a+h)-f(a))/h)$ provided that $f$ is continuous at $a$ (or in other words, that $\lim_{x \to a}(f(x))=f(a)$).
One gripe that I have with this definition is that the possible output values of the limit are $\mathbb{R} \cup \{\infty, -\infty\}$, meaning that the possible values of the derivative are also $\mathbb{R} \cup \{\infty, -\infty\}$. I don't think it makes much sense for the derivative to equal infinity at a point where the tangent line is vertical for a single point on the curve. The derivative is supposed to measure the slope of the tangent line, and a vertical tangent line has no defined slope. One such example is the function $f(x) = \sqrt{x}$, which is continuous at $0$, and whose derivative at $0$ is equal to $\infty$. Both of those can be proven by using delta-epsilon techniques.
What is needed is a "cast" from the output value of the limit to the real numbers before assigning it to the derivative. For example, we should say that if $\lim_{h \to 0}((f(a+h)-f(a))/h) \in \mathbb{R}$, then $f'(a) = \lim_{h \to 0}((f(a+h)-f(a))/h)$ provided that $\lim_{x \to a}(f(x))=f(a)$.