# Perfect set of irrationals

I solved the following exercise:

Let $\{r_1,r_2,r_3,\dots \}$ be an enumeration of the rational numbers and for each $n \in \mathbb N$ let $\varepsilon_n = 1/2^n$. Define $O = \bigcup_{n=1}^\infty V_{\varepsilon_n}(r_n)$ and let $F=O^c$.

(a) Argue that $F$ is closed and nonempty consisting only of irrational numbers.

(b) Does $F$ contain any nonempty open intervals? Is $F$ totally disconnected?

However I am now stuck with (c):

(c) Is it possible to know whether $F$ is perfect? If not, can we modify this construction to produce a nonempty perfect set of irrational numbers?

I tried and I can't do it but I would really like to know the answer. Of course it is not possible to know that $F$ is perfect because it might contain isolated points but can there be a perfect set of irrational numbers?

• What are $V_{\varepsilon_n}(r_n)$ and $O^c$ ? – Gabriel Romon Jan 5 '14 at 13:30
• Do you means all number in the set must be irrational, or you means the set must contains all irrational number? – Gina Jan 5 '14 at 13:32
• Gina: he means the second one. Gabriel: $V_{\epsilon_n}(r_n)$ and $O^c$ are baby-Rudin notation for "the ball of radius $\epsilon_n$ about $r_n$" and "the set-compliment of $O$" respectively. – Omnomnomnom Jan 5 '14 at 13:41
• @Gina I mean that it should be a possibly strict subset of the irrational numbers. – newb Jan 5 '14 at 14:56
• @GabrielR. It denotes the open ball around $r_n$ or radius $\varepsilon_n$ and the other is the complement of $O$ in $\mathbb R$. – newb Jan 5 '14 at 14:57

Note however, that by our construction, that border must be rational, which means that it is contained by some other neighborhood in $O$.
• The boundary points of all the $V_{\varepsilon_n}(r_n)$ are rational, so if two such sets have a common boundary point, that does not belong to $F$. The problematic situation is when the the boundary points of the $V_{\varepsilon_n}(r_n)$ approach an irrational $x$ from both sides, can it happen that $O$ contains $(x-\delta,x)\cup (x,x+\delta)$ for some $\delta > 0$? – Daniel Fischer Jan 5 '14 at 14:15