3
$\begingroup$

Let $P: (OP_{X}) \to (Sets)$ be a sheaf, where $(OP_X)$ is just the category of open sets of a topological space $X$. Define $P_x$ to be the stalk of $x \in X$. Then we can define the Etale space to be $Et(P):=\amalg_{x \in X} P_x$, and the topology is given by a basis $\{U_s| U \ open\ in X\}$, where $U_s = \{S_x|x \in U\}$($S_x$ is the germ of $S$ at $x$). There is a natural projection $\pi: Et(p) \to X$, which is continuous and surjective.

Now the claim is that there is a natural continuous map $g: X \to Et(p)$ such that $\pi \circ g = id_{X}$.

My approach is to associate with each point $x \in X$ a section $s \in P(X)$, since $X$ is a natural open set containing $x$. However this choice of section seems to be too arbitrary to be continuous.

The reason that I ask about this is that the existence of a global section ensures that we can define $P^{+}$ the sheaf associated to $P$ because $P^{+}(U)$ is defined to be the set of all continuous sections on $U$ and the restriction of a global section onto each open subset ensures that $F^{+}(U)$ is not empty.

This answer gives a more precise account of this construction

Update: Inspired by the given link, given $s \in P(X)$ we can consider directly the map $g:X \to Et(P)$, $g(x) = s_x$. Now I want to show that it is continuous. $g(x) \in U_{s'}$ iff there is an open set $U'$ containing $x$ such that $r^{X}_{U'}s = r^{U}_{U'}s'$. What then?

$\endgroup$
3
  • $\begingroup$ "natural"? Associated to each global section of the sheaf there such a map $g$. Indeed such maps $g$ correspond exactly to the global sections of the sheaf (so when there are no global sections, there are no such $g$). $\endgroup$ Commented Jul 4, 2019 at 2:52
  • $\begingroup$ I added something in the question. If a global section does not exist then $F^{+}$ cannot always be defined which seems wrong to me. @LordSharktheUnknown $\endgroup$
    – Keith
    Commented Jul 4, 2019 at 3:16
  • 1
    $\begingroup$ The definition of the sheaf $P^+(U)$ uses sections on every open subset $U$, not necessarily global sections. If for any $U$ there is no sections on $U$, well, then you have the empty presheaf and its sheafification will be the empty sheaf. But if $P$ has no global section, it might have sections on smaller $U$ and the sheafification won't necessarily be empty. $\endgroup$
    – Roland
    Commented Jul 4, 2019 at 8:08

0

You must log in to answer this question.