Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have the following metric space: The set $X$ of all sequences with members from the set $\{1,2,\ldots, n\}$, together with the metric $$d(x,y)=\frac{1}{\min\{j\in\mathbb{N}:x_j\ne y_j\}}.$$ I wish to prove two things about this space:

1) $X$ is compact.

2) If $T:X\to X$ is defined by $Tx_n=x_{n+1}$ then $T$ is continuous.

Its many years since I studied topology so if someone can help me even a little bit I will be very grateful. Thanks.

share|cite|improve this question
Maybe we should write $(Tx)_n$? I was confused at first. – Dylan Moreland Jul 8 '12 at 6:16
up vote 7 down vote accepted

Hint for 1: Note that a metric space is compact iff every sequence has a convergent sub-sequence. Let $(y_k)$ be a sequence in $X$. Since there are only finitely many choices for the first term of an element of $(y_k)$, there is some sub-sequence $(y_k^1)$ of $(y_k)$ such that all elements have the same first term. Since there are only finitely many choices for the second term, there is some sub-sequence $(y_k^2)$ of $(y_k^1)$ such that all elements have the same first term. Continue in this manner, and show that the sub-sequence $(y_k^k)$ (often called the diagonal sub-sequence) is convergent.

Hint for 2: Let $(x_j),(y_j)\in X$. Note that $\min\{j\in \mathbb N: x_j\neq y_j\}=\min\{j\in \mathbb N: x_{j+1}\neq y_{j+1}\}+1$, so $$\frac{1}{d((x_j),(y_j))}=\frac{1}{d(T(x_j),T(y_j))}+1$$ and do some rearranging.

share|cite|improve this answer
Do you mean +1 in Hint 2? For $(x_j)=1,2,3\cdots$ and $(y_j)=1,2,4\cdots$, the min on the left is 3 while the min on the right is 2. Or am I missing something here? – Shahab Jul 8 '12 at 6:44
@Shahab: Yes, it should be $+1$. – Brian M. Scott Jul 8 '12 at 6:50

To prove that $X$ is compact, it suffice to show that it is totally bounded and complete.

To show that it is totally bounded, let $\epsilon\gt 0$. Let $m\gt 0$ be such that $\frac{1}{m}\lt\epsilon$. Now, there are only $n^m$ possible finite sequences of length $m$ with values in $\{1,2,\ldots,n\}$. For each such sequence $s$, let $\mathcal{O}_s$ consist of all elements of $X$ that agree with $s$ in the first $m$ points. The set is open, since for every $(x_n)\in \mathcal{O}_s$, any sequence within $\frac{1}{m}$ of $(x_n)$ must agree with $x_n$ on the first $m$ terms, and hence have $s$ as its initial term. Moreover, the distance between any two elements of the set is less than $\epsilon$, so the radius of $\mathcal{O}_s$ is less than $\epsilon$. Since $X$ is the union of the $\mathcal{O}_s$, we see that $X$ is totally bounded.

Now suppose that $(x_n)_m$ is a Cauchy sequence. Then for every $k\in\mathbb{N}$ there exists $M(k)\gt 0$ such that if $j,\ell\geq M(k)$, then $d((x_n)_i,(x_n)_j)\lt \frac{1}{k}$. That means that $(x_n)_i$ and $(x_n)_j$ agree in the first $k$ terms. Define the sequence $(y_n)$ by letting $y_k$ be the common $k$th term to all sequences with index greater than $M(k)$. Show that $(x_n)_m\to (y_n)$.

Thus, $X$ is compact.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.