Let $X$ be a topological space. $Y$ is a discrete subset of $X$ if it has a discrete topology induced by the topology of $X$. This is equivalent to the fact that for every $y\in Y$ there is an open $U\subset X$ such that $U\cap Y=\{y\}$. Consider a stronger condition: there is a collection $\{U_y\}_{y\in Y}$ of mutualy disjoint open sets, such that $y\in U_y$. I wonder if the two are equivalent for nice spaces, but my question today is following:

Is it true that if $X$ is locally compact and not compact, there is a discrete (in the stronger sense) set $Y\subset X$, such that $\overline{Y}$ is not compact?

In my specific case $X$ is locally compact, but this condition is not necessary in order to pose the question. If $X$ is metrizable and separable, it has a metrizable compactification, and $Y$ can be chosen as a sequence that converges in a conrolled way to an infinite element. If $X$ is locally compact and $\sigma$-compact, then there is a sequence of compacts $\{K_n\}$ such that $K_n\subset int~K_{n+1}$, and taking $Y$ to be the sequence of arbitrary $y_n\in int~K_{n+1}\backslash K_{n}$ will work.

I also have a reference request about existence of an infinite discrete subset in any infinite Hausdorff space. This follows from the result in this article, but I was not able to find a more direct reference. I found some research on this topic, but not this explicit statement.

  • 1
    $\begingroup$ To last your question there exists a direct argument: if a Hausdorff space $X$ is not discrete, then take a non-isolated point $x\in X$ and by induction construct a decreasing sequence $(U_n)_{n\in\omega}$ of neighborhoods of $x$ such that for every $n$ the set $U_{n+1}$ is disjoint with a neighborhood $V_n\subset U_n$ of some point $x_n\in U_n$. Then $\{x_n\}_{n\in\omega}$ will be a requred strongly discrete set in $X$. $\endgroup$ – Taras Banakh Apr 7 '18 at 18:21
  • $\begingroup$ If your space is locally compact and $T_2$ then it's completely regular; furthermore assuming it's not countably compact, you should be able to thin-out, a countable set without accumulation point, to get your desired discrete set (in the strong sense.) If the space is countably compact, I suspect the answer does not easily follow. $\endgroup$ – Not Mike Apr 7 '18 at 19:31

The answer to this question is negative and can be obtained with the help of weak P-points in compact spaces with countable cellularity. A non-isolated point $p$ of a topological space $X$ is a weak $P$-point if $p$ is not an accumulation point of a countable set $C\subset X\setminus\{p\}$.

If $p$ is a weak P-point in a compact space $X$ with countable cellularity, then $X\setminus\{p\}$ is a locally compact space such that any strongly discrete subset is countable and hence has compact closure in $X\setminus\{p\}$.

It seems that examples of countably cellular spaces with weak P-points can be found using this result of Dow.

  • $\begingroup$ Thank you! I have to admit that I did not know the definitions of neither countable celluarity, nor P-points. I guess if I knew, I wouldn't have asked the question. Yet another evidence of how numb the intuition is sometimes, and how broad the mathematics is. $\endgroup$ – erz Apr 7 '18 at 23:18
  • $\begingroup$ @erz In fact, your intuition was not very far from the truth: by an old (1988) result of Tkachuk a Tychonoff space $X$ is compact if each discrete subspace of has compact closure. $\endgroup$ – Taras Banakh Apr 9 '18 at 7:34
  • $\begingroup$ So these two notions of discreteness are different after all, right? $\endgroup$ – erz Apr 9 '18 at 7:55

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.