Let $I = [0,1]$ and $I^{\omega}$ be the countable product of unit closed interval I, where each $I$ is given the subspace topology of $R$ in the usual topology.

I am trying to show that $I^{\omega}$ is not compact in the box and the uniform topology by showing that there is an open cover without a finite subcover.

In the box topology case, consider the open cover,

${∏_{n∈N} I_n :I_n∈{[0,2/3 ),(1/3 ,1]} }$

It is intuitively clear that there is no finite subcover of this open cover, yet I cannot state it out in words.

For the uniform topology case, I cannot think of such an open cover. Can somebody help me?

  • $\begingroup$ Is $[0,1]$ on its own compact? $\endgroup$ – graydad Nov 25 '14 at 3:14
  • $\begingroup$ Sorry I forgot to mention that I is given the subspace topology of usual R. $\endgroup$ – takecare Nov 25 '14 at 3:19
  • $\begingroup$ Tychonoff is with respect to the product topology. $\endgroup$ – takecare Nov 25 '14 at 3:21
  • $\begingroup$ Ah; right you are. My mistake! $\endgroup$ – graydad Nov 25 '14 at 3:22

Let’s begin by completing your idea for the box topology. Let $U=\left[0,\frac23\right)^\omega$; this is open in the box topology. For $n\in\omega$ let

$$V_n=\left\{\langle x_k:k\in\omega\rangle\in I^\omega:x\in\left(\frac13,1\right]\right\}\;;$$

each $V_n$ is open in the box topology — indeed, even in the ordinary product topology — and $\mathscr{V}=\{U\}\cup\{V_n:n\in\omega\}$ is an open cover of $I^\omega$. Suppose that $\mathscr{V}_0\subseteq\mathscr{V}$ is finite. Let $F=\{n\in\omega:V_n\in\mathscr{V}_0\}$. Define a point $x=\langle x_n:n\in\omega\rangle\in I^\omega$ as follows:

$$x_n=\begin{cases} 0,&\text{if }n\in F\\ 1,&\text{otherwise}\;. \end{cases}$$

Then $$x\notin U\cup\bigcup_{n\in F}V_n\supseteq\bigcup\mathscr{V}_0\;,$$

so $\mathscr{V}_0$ doesn’t cover $I^\omega$, and $\mathscr{V}$ has no finite subcover.

For the uniform topology I think that another approach is easier. For $n\in\omega$ define the point $x^{(n)}=\langle x_k^{(n)}:k\in\omega\rangle\in I^\omega$ by

$$x_k^{(n)}=\begin{cases} 1,&\text{if }k=n\\ 0,&\text{otherwise}\;. \end{cases}$$

It’s not hard to show that $A=\{x^{(n)}:n\in\omega\}$ is an infinite subset of $I^\omega$ with no limit point in the uniform topology, which is enough to show that $I^\omega$ is not compact in the uniform topology. If you want to use open covers, for each $n\in\omega$ let $B_n$ be the open $\frac12$-ball centred at $x^{(n)}$. (In other words, $B_n=\prod_{n\in\omega}I_k$, where $I_k=\left[0,\frac12\right)$ if $k\ne n$, and $I_n=\left(\frac12,1\right]$.) Then show that $A$ is closed, and let $U=I^\omega\setminus A$; $\mathscr{U}=\{U\}\cup\{B_n:n\in\omega\}$ is then an open cover of $I^\omega$ with no finite subcover, since for each $n\in\omega$ the only member of $\mathscr{U}$ containing $x^{(n)}$ is $B_n$.


One way to do it: consider the subset $A = \{0,1\}^\omega$ consisting of sequences whose entries are all either 0 or 1. Show that in each topology, $A$ is closed. For each $x \in A$, find an open set $U_x$ with $U_x \cap A = \{x\}$. Show that the open cover $\{A^c\} \cup \{U_x : x \in A\}$ has no finite subcover.


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.