Relevant portion from Rudin's PMA

Theorem 4.20 continued

To make explicit the argument that $f$ is not uniformly continuous, I tried the following. I wanted to ask if it looked alright.

Consider $$f(x) = \frac{1}{x - x_0}$$ Here, $x_0$ is a limit point of $E$ and $x_0 \notin E$. We claim that is not uniformly continuous on $E$. Take $\epsilon = 1$. We will show that given $\delta > 0$, there exists $t_\delta, x_\delta \in E$ with $|t_\delta - x_\delta|< \delta$ such that $|f(x_\delta) - f(t_\delta)| \geq 1$.

Choose $x_\delta$ such that $|x_\delta - x_0| < \delta$. Note that the $\delta$-neighbourhood of $x_\delta$ is a neighbourhood of $x_0$, so it contains points arbitrarily close to $x_0$. Thus there exists $t_\delta$ with $|x_\delta - t_\delta| < \delta$ such that $|t_\delta - x_0| < \frac{1}{1 + k}$ where $k = \frac{1}{|x_\delta - x_0|}$. Then $$|f(x_\delta) - f(t_\delta)| \geq |f(t_\delta)| - |f(x_\delta)| = \left|\frac{1}{x_0 - t_\delta}\right| - \left|\frac{1}{x_0 - x_\delta}\right| \geq 1 $$


Your proof is fine. However, I would not have used the letter $k$ in this context, for psychological (not logical) reasons. It's just that $k$ usually represents an integer, but here it can be any positive real.


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.