I am self-studying the book, Walter Rudin, Principles of Mathematical Analysis, and I am doing the Exercise 2.26.

enter image description here

As mentioned in the hint, Exercise 23 suggests: every separable metric space has a countable base. According to the definitions, a metric space is separable if it contains a countable dense subset. A base $\{V_\alpha\}$ is a collection of sets such that every open set in X is the union of a subcollection in ${V_\alpha}$.

Exercise 24 suggests that, if in a metric space $X$, every infinite subset has a limit point, then $X$ is separable.

Following the hint, $X$ is separable, which means $X$ has a countable (infinite but countably many) base.

To prove $X$ is compact, we need to show for any open cover $\{G_\alpha\}$, there exists a finite subcover $\{G_1, G_2, \cdots, G_n\}$.

If I understand correctly, the hint and solution suggest first finding a countable subcover and then finding a finite subcover.

However, I do not see why "It follows that every open cover of $X$ has a countable subcover $\{G_n\}_{n=1}, n=1, 2, 3, \cdots$."

This was not proved in the hint or solution. Maybe trivial, but I can't see it at this moment. Can anyone give a hint? Thanks.


Let $(U_i)_{i\in \mathbb{N}}$ a countabla basis, and consider an open covering $(V_j)_{j\in J}$, for every $x\in X$, there exists $j_x\in J$ such that $x\in V_{j_x}$. There exists $i_x\in \mathbb{N}$ such that $x\in U_{i_x}\subset V_{j_x}$ since $(U_i)_{i\in\mathbb{N}}$ is a basis of the topology. Consider the map $f:X\rightarrow\mathbb{N}$ defined by $f(x)=i_x$ and the map $g:\mathbb{{N}}\rightarrow J$ defined by $g(n)=j_y$ where we choose one $y\in X$ and $f(y)=n$. $X$ is covered by $V_{g(n)}$. To see this, let $x\in X$, suppose that $g(i_x)=j_y$, we have $g(y)=i_x$. This implies that $U_{i_x}=U_{i_y}$ and $x\in U_{i_x}=U_{i_y}\subset V_{j_y}=V_{g(n)}$.

  • $\begingroup$ Thanks. This enables me to continue with the rest of the proof. $\endgroup$ – Mou Apr 21 '19 at 22:44
  • $\begingroup$ Hi Tsemo, I somehow lost the big picture of the proof. Why does it lead to a contradiction that z cannot be a limit point of {x_m}? Thank you. $\endgroup$ – Mou Apr 22 '19 at 1:38
  • $\begingroup$ By reading another post, I now understand, if the post is correct, that B(z) contains only finite elements of E, and thus it could not be a limit point. I was stuck at the provided solution because it says z is not a limit point of {x_m}. This is always true no matter if it means {x_m} for some m, or {x_m: m>=n}. Indeed, we should determine if z is a limit point of E (not {x_m} in either sense). $\endgroup$ – Mou Apr 22 '19 at 2:00

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.