According to Berkeley Physics Course, Volume 1 Mechanics,
The time derivative of a vector valued function can be derived from the formula: $$ \mathbf{r}(t) = r(t)\mathbf{\hat{r}}(t) $$ where the scalar $r(t)$ is the length of the vector and $\mathbf{\hat{r}}(t)$ is a vector of unit length in the direction of $\mathbf{r}$. The derivative of $\mathbf{r}(t)$ is defined as: $$ \begin{equation} \tag{1} \frac{d\mathbf{r}}{dt} = \frac{d}{dt}[r(t)\mathbf{\hat{r}}(t)] = \lim_{\Delta t \to 0}\frac{r(t+\Delta t)\mathbf{\hat{r}}(t+\Delta t) - r(t)\mathbf{\hat{r}}(t)}{\Delta t} \end{equation} $$ Now we use the Taylor series expansion for each of the factors $r(t+\Delta t)$ and $\mathbf{\hat{r}}(t+\Delta t) $ and we only retain the first two terms in the series expansions of $r(t+\Delta t)$ and $\mathbf{\hat{r}}(t+\Delta t)$. The general Taylor expansion series is:
$$ \begin{equation} \tag{2} f(x) = f(x_{0}) + (x-x_0)\left[ \frac{df(x)}{dx}\right]_{x=x_0} + \frac{1}{2}(x-x_{0})^{2} \left[ \frac {d^{2}f(x)}{dx^{2}} \right]_{x=x_0} + \cdots \end{equation} $$
After substituting the first two terms of the Taylor expansion series(2) into the numerator of equation $(1)$, we obtain $$ \begin{equation} \tag{3} \left[r(t)+\frac{dr}{dt} \Delta t\right] \left[\mathbf{\hat{r}}(t) + \frac {d\mathbf{\hat{r}}}{dt} \Delta t\right] - r(t)\mathbf{\hat{r}}(t) = \Delta t \left(\frac{dr}{dt}\mathbf{{\hat{r}}} + r\frac{d\mathbf{\hat{r}}}{dt} \right) + (\Delta t)^{2}\left(\frac{dr}{dt}\frac{d\mathbf{\hat{r}}}{dt}\right) \end{equation} $$
Substituting $(3)$ in $(1)$, we obtain: $$ \begin{equation} \tag{4} \frac{d\mathbf{r}}{dt} = \frac{dr}{dt}\mathbf{\hat{r}} + r\frac{d\mathbf{\hat{r}}}{dt} \end{equation} $$
Equation $(4)$ is the general formula for the derivative of the product of a scalar function and a vector valued function.
My question is that why do we prove this formula as an estimated value? Why can't we directly use the product rule for differentiation here?