I am taking a course on Manifolds, and we glossed over the definition of a topological manifold without much discussion. The thing I am specifically unsure about, is what the basis of a manifold is.
From previous intuition, I was thinking about a basis akin to that of a vector space. But we do not have global coordinates and so clearly this is the wrong interpretation.
Doing a bit of research, it seems that the basis is a topological basis. But I do not fully understand that in this context. Is our basis the collection of all points in our manifold (which seems unlikely because then it would rarely be countable?), or is it the collection of open sets?
If it is the collection of open sets, then will each open set be contained in a chart of the maximal atlas? Any elaboration on the nature of this basis would be very helpful in my understanding.
A base (basis is usually reserved for algebraic uses of the word) for $X$ is set $\mathcal{B}$ of open subsets of $X$ such that for every open set $O$ and every $x \in O$ we have some $B \in \mathcal{B}$ such that $x \in B \subseteq O$. It turns out that this notion is quite handy to discuss topological properties: continuity of a function can be determined by looking at base elements e.g. and in a manifold the locally Euclidean (homeomorphic to $\Bbb R^n$ for some $n$) open sets form a base by definition. In many cases restrictions are put on the size of a base, especially spaces with countable bases are a popular subclass of spaces: this implies that the space is Lindelöf, separable etc. and makes it more amenable to analysis techniques (we can use countable series, countable covers by locally Euclidean neighbourhoods etc.) So very often (in many texts) manifolds are assumed to have a countable base (as all $\Bbb R^n$ have). A maximal atlas is always a base and if the manifold has a countable base, we can reduce any atlas to some countable atlas (which makes it more manageable), by standard theorems in general topology.