If I have a manifold $M$ of $\dim M=m$ that has an atlas with one map $(M; \varphi)$. If we denote the components of $\varphi$ with $x^1,\dots, x^m$.
Does the following operation make sense?
$\frac{\partial}{\partial x^l}|_p$ is an element of $T_pM$ so it is aderivation that can act on smooth functions over the manifolds such as $x^l$
$ \frac{\partial }{\partial x^l}|_p x^j =\frac{\partial x^j}{\partial x^l}|_p=\delta^j_l $
Why or why not? Can I really see this a elementary derivation, or are there intermediate steps that justify this is true in the manifold case? I am a bit confused because $x^j$ is a function and a variable at the same time when I do that partial derivative and I am working as if M was $R^m$
Let $f :M \to \mathbb{R}$ a smooth function. If we denote the components of $\mathbb{R}^m$ with $r^1 \dots r^m$, by definition $$ \frac{\partial}{\partial x^j}\Bigg|_p(f) = \frac{\partial}{\partial r^j}\Bigg|_{\varphi(p)}(f \circ \varphi^{-1}) $$ where the RHS is a derivation with respect to Euclidean coordinates. So in your case you can compute $$ \frac{\partial}{\partial x^j}\Bigg|_p(x^i) = \frac{\partial}{\partial r^j}\Bigg|_{\varphi(p)}(x^i \circ \varphi^{-1} ) = \frac{\partial}{\partial r^j}\Bigg|_{\varphi(p)}(r^i \circ \varphi \circ \varphi^{-1} ) = \frac{\partial r^i}{\partial r^j}\Bigg|_{\varphi(p)} = \delta^i_j. $$
This reflects that manifolds are modeled as spaces which behave locally as the Euclidean space (at least when you make first order computations).