I am trying to prove the following:
$K \subset U \subset \mathbb R^n$, where $K$ is compact and $U$ open, then there is $\epsilon > 0$, such that $\forall u \in \mathbb R^n \setminus U, \forall k \in K, d(k,u) \ge \epsilon$, where $d$ is the usual metric on $\mathbb R^n$.
I was trying to prove it by contradiction. Assume that $\forall n, \exists k_n \in K, u_n \in \mathbb R^n \setminus U: d(k_n,u_n) < \frac{1}{n}$, then clearly $k_n-u_n \to 0$.
$K$ is compact and thereby constrained, thus $(k_n)_{n \in \mathbb N}$ has a convergent sub-sequence $(k_{n_m})_{m \in \mathbb N}$ which converges in $K$ because $K$ is closed, thus $ (k_{n_m})_{m \in \mathbb N} \to k \in K$.
It means that $u_{n_m} = u_{n_m} - k_{n_m} + k_{n_m} \to k \in \mathbb R^n \setminus U $ , because $\mathbb R^n \setminus U$ is also closed.
But then $k \in K \subset U$ and $k \in \mathbb R^n \setminus U$, so we have got a contradiction.
Have I proved it, or there is an error somewhere?