Let $X$ be a Hausdorff, locally compact but non-compact topological space. If the (Alexandroff) one-point compactification is connected, can $X$ have compact connected components?
-
Do you mean all connected components are compact $;$ or $;$ there is a compact component ? – Stefan Hamcke Nov 04 '13 at 17:10
2 Answers
I think I proved the following
Lemma
Let $X$ be a Hausdorff space and $C \subset X$ have a compact neighbourhood $K$. Then $C$ is a component of $X$ if and only if $C$ is a component of $K$.
in this answer.
For the present problem this implies a negative answer, since a compact set in a locally compact Hausdorff space has a compact neighbourhood. Let $C$ be a compact component of $X$, and let $K$ be some compact neighbourhood of $C$ in $X$. Applying the lemma one way then shows that $C$ is a component of $K$, and then applying it the other way shows that $C$ is a component of the compactification, since $K$ remains a compact neighbourhood of $C$.

- 6,835
-
+1. That Lemma seems correct (after going through it) and it is quite powerful :) Thanks again for pointing out my mistake in my answer below. – Stefan Hamcke Feb 08 '14 at 23:16
-
@StefanHamcke: Thanks for confirming that. I wasn't too confident since hardly anyone else seemed to even have looked at it. – Niels J. Diepeveen Feb 09 '14 at 01:13
Well, no. Assume $X=Y\sqcup C$, where $C$ is compact. Then the 1PC is $X\cup\{\infty\}$ with the topology: $U$ is open iff $U\subset X$ is open in the previous topology or $\infty\in U$ and $U^c$ is compact. Then $C = (U\cup\{\infty\})^c$ is compact, and thus $U\cup\{\infty\}$ is open. $C$ is also open (it was open in $X$). Thus the 1PC of $X$ is not connected.

- 29,795
-
-
2@Stefan: It certainly isn’t true in general for locally compact, non-compact Hausdorff spaces; just look at the Cantor set minus a point, for instance. – Brian M. Scott Nov 04 '13 at 18:14
-
-
@iago Of course it can. It will also be closed, of course, but this doesn't exclude it from being open. Indeed, it is open by my definition of $X$ as a disjoint union. – Daniel Robert-Nicoud Nov 03 '19 at 10:59