The Topologist's Sine Curve is well-known:

The set $S = \{(0,0)\} \cup \{(x, \sin(1/x))\ |\ x \in \left]0,1\right]\}$, as a subspace of $\mathbb{R}^2$, is connected but not path-connected.

An intuitive reason is that no path from $S - \{(0,0)\}$ can reach $(0,0)$ in a finite amount of time. However, what if I shrink the distances involved? That is, consider $$ S' = \{(0,0)\} \cup \{(x, x\sin(1/x))\ |\ x \in \left]0,1\right]\}. $$

Is $S'$ path-connected?


Let $$f:[0,1]\to\Bbb R:x\mapsto\begin{cases}x\sin\frac1x,&\text{if }0<x\le 1\\\\0,&\text{if }x=0\;;\end{cases}$$

then $f$ itself is already a path connecting any two points of $S'$.

  • $\begingroup$ Indeed, as $\lim_{x\to 0} x\sin{(\frac{1}{x})} = 0 = f(0)$ $\endgroup$ – Rustyn Mar 17 '13 at 5:09
  • $\begingroup$ @Rustyn: Exactly. (But I figured that that was pretty well known.) $\endgroup$ – Brian M. Scott Mar 17 '13 at 5:10
  • $\begingroup$ +1 nice answer@ Brian sir $\endgroup$ – user525416 May 19 '18 at 1:13

I should point out some subtlety here. The problem is never about infinite length of these curves. Remember that we have space filling (continuous) curves into $R^2$. So, why is it possible to turn the shrinking sine cure into a path, but the same is not possible for the usual sine curve?

The real answer is the non-/existence of the limit at zero. In both cases we already have a curve defined from $(0,1]$ into your space. In one case we can assign a value at $0$, and extend the function's domain continuously to the closed $[0,1]$, while in the other we cannot, since there is no limit at zero.

When can we extend a continuous function from $(0,1]$ to all of $[0,1]$? Precisely, if and only if the (original) function is uniformly continuous.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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