Computing Directional Derivatives of Vector Fields

61 Views Asked by At

So in the context of differential geometry and covariant derivatives, I've seen the following formula for computing directional derivatives of a vector field $\mathbf X$.

$$D_{\mathbf V} \mathbf X = (\mathbf X(\mathbf{\alpha}(t)))'(0)$$

where $\mathbf \alpha$ is a curve with $\alpha(0) = P$ and $\alpha ' (0) = \mathbf V$.

But how does one use this in practice? Using the chain rule, you'd want the derivative of $\alpha$ times the derivative of $\mathbf X$ of $\alpha$. But if you use the dot product, that would give you a scalar result. It seems like it should be a vector result, since the derivative of a vector function has the same amount of dimensions as the original.

So how do you use this formula in practice?