0
$\begingroup$

I am studying the theorem "if $f:(X,d_X)\to (Y,d_Y)$ is continuous and $X$ is compact, then $f$ is uniformly continuous." I am not looking for a proof, but I have an argument against any attempt at a direct proof (which was my original attempt at proving this statement).

Why a direct proof fails: Since $X$ is open, for each $x\in X$ there is a $\delta_x$ such that the open ball $B(x,\delta_x)$ is contained in $X$. Hence, I can write $X=\bigcup_{x\in X}B(x,\delta_x)$. Moreover, since once we are given a particular $\delta_x$ any smaller radii will suffice, I can cover $X$ in such a way that $\inf\lbrace \delta_x:x\in X\rbrace=0$ by choosing arbitrarily small balls. If given $\varepsilon>0$ how can I ever expect to find a $\delta>0$ such that $$d_X(x,y)<\delta\Rightarrow d_Y(f(x),f(y))<\varepsilon\,\,\,\,\forall x,y\in X$$ if the balls on which $f$ is continuous can be of arbitrarily small radii?

$\endgroup$
  • $\begingroup$ If $X$ has only finitely many elements, then it's not true that you can achieve $\inf\{\delta_x : x \in X\} = 0$. Compact sets are in some sense "almost finite" because you can reduce an infinite cover (where the inf might be zero) to a finite one (where it cannot be). $\endgroup$ – Bungo Sep 5 '14 at 4:17
  • $\begingroup$ First of all, $X$ cannot be open because it is compact. $\endgroup$ – Jango Dec 28 '17 at 0:28
3
$\begingroup$

How is this not "direct"?

Let $\epsilon > 0$ be given. Since $f$ is continuous, for each $x \in X$ there is $\delta_x$ such that $d_X(x,y) < \delta$ implies $d_Y(f(x),f(y)) < \epsilon/3$. The balls $B(x,\delta_x/2)$ cover $X$, therefore so does some finite subcover $\{B(x_i,\delta_{x_i}/2): i=1\ldots n\}$. Take $\delta = \min(\delta_{x_i}/2: i=1\ldots n)$. Suppose $d(x,y) < \delta$. There is some $i$ such that $d(x, x_i) < \delta_{x_i}/2$, and $d(y,x_i) \le d(y,x) + d(x,x_i) < \delta_{x_i}$. Therefore $d(f(x),f(y)) \le d(f(x),f(x_i)) + d(f(y),f(x_i)) < \epsilon$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.