Can't figure out what's wrong with my proof

83 Views Asked by At

I have to decide if it possible to find a set $A\subset \mathbb{R}$ such that:

$A$ is not connected nor compact but it is complete.

At first, I thought it wasn't possible, and made the following "proof":

Proof: Given that $A$ is not connected, it is possible to find $B$, $C$ subsets of $\mathbb{R}$, disjoint and open such that $A=B\cup C$. Using the "representation theorem for open sets on the real line" (math analysis, Apostol), $B$ and $C$ are both a countable union of disjoint open intervals, so $A=\bigcup_{k=1}^{\infty}(a_k,b_k)$ where one $a_j$ and $b_j$ can be $-\infty$ and $+\infty$, respectively.

So if we take one interval with a finite endpoint, say, $(a_2,b_2)$ ($a_2$ finite), we can construct a sequence converging to $a_2$, so that sequence is cauchy, but doesn't converge to a point belonging to $A$. $\blacksquare$

But a classmate could find an example: $\mathbb{Z}$. Another one: $\bigcup_{k=0}^{\infty}[2k,2k+1]$.

So, what's wrong with my proof? Any help is highly appreciated. Thanks and regards.

2

There are 2 best solutions below

0
On BEST ANSWER

You need to be careful when you write $A$ as a disjoint union of sets. We have that $A = B\cup C$ where $B$ and $C$ are open in $A$, but they don't have to be open in $\mathbb{R}$. Consider $A = [0,1) \cup [2,3]$.

As a hint to finding a disconnected, noncompact, complete set, I would consider an interval of the form $(-\infty, a]$ for some $a \in \mathbb{R}$. There is a way to construct a set with the desired properties from an interval of this form.

0
On

If $A$ is not connected you can find $B,C\subset A$ which are open in $\boldsymbol A$ such that $B\cap C=\emptyset$. $B$ and $C$ do not have to be open in $\Bbb R$.