Questions about étale space construction of sheafification

78 Views Asked by At

I'm trying to wrap my head around sheafification via the étale space construction from, say, Hartshorne's or Liu's books on algebraic geometry (I guess this is a construction of Bourbaki?). In these constructions, for a presheaf $\mathcal{F}$ on a topological space $X$ with stalks $\mathcal{F}_x$ at $x$, one defines the sheafification $\mathcal{F}^{\dagger}$ to have sections $$\mathcal{F}^{\dagger}(U) = \left\{f: U \to \amalg_{x \in U} \mathcal{F}_x\, |\, \star \right\}$$ where $U \subseteq X$ is open and $\star$ is the condition that for each $x \in U$ there exists an open neighborhood $x \in V \subseteq U$ and a section $s \in \mathcal{F}(V)$ verifying $f(y) = s_y$ for all $y \in V$. That is, that $f$ doesn't pick germs arbitrarily, but picks out "compatible local sections".

In thinking of how I would've intuited a definition of $\mathcal{F}^{\dagger}(U)$ as a limit over some diagram of the $\mathcal{F}_x$'s, this notion in terms of functions struck me as odd until I realized that, in Set at least, $$\prod_{x \in U} \mathcal{F}_x \cong \left\{f: U \to \coprod_{x \in U} \mathcal{F}_x \right\}.$$

Q1. This implies that I can construct sections of a sheaf, not just here but in general as some kind of limit of the stalks, but I'm unsure what the transition maps should be. I am familiar with a construction of the sections over an open $U$ relative to an open cover, but I'm not sure how to take that "to the stalks". How can one construct $\mathbf{\mathcal{F}^{\dagger}(U)}$ as a limit over a diagram of stalks and homomorphisms?