3
$\begingroup$

Let $X\subset \mathcal{B}(\mathbb{R},\mathbb{R})$ be the subset of bounded functions $f:\mathbb{R}\to\mathbb{R}$ such that are discontinuous in all points. Prove that $X$ is not open (with usual metric of supremum).

I think that I found an example of a function $f\in X$ such that for every $\epsilon>0$ there exist $g\in B(f,\epsilon)\cap X^c$.

Let $f$ defined by:

$$f(x)=\begin{cases} 1\ ; \text{if } x\in[-1,1]^c\cap\mathbb{Q}\\ x\ ; \text{if } x\in[-1,1]\cap\mathbb{Q}\setminus\{0\}\\ 1/2\ ; \text{if } x=0\\ -x\ ; \text{if } x\in[-1,1]\cap\mathbb{Q}^c\\ -1\ ; \text{if } x\in[-1,1]^c\cap\mathbb{Q}^c\\ \end{cases}$$

$f$ clearly be in $X$ and for each $\epsilon>0$ I can define $g$ by:

$$g(x)=\begin{cases} f(x)\ ; \text{if } x\in\{0\}\cup(-\epsilon,\epsilon)^c\\ 0\ ; \text{if } x\in(-\epsilon,\epsilon)\setminus\{0\}\\ \end{cases}$$

By a "draw" I think that $g\in B(f,1.1\epsilon)$ and $g$ is continuous in an interval then $g\in X^c$. Therefore $X$ is not open.

Am I right?

Do you know another way to prove this fact?

$\endgroup$
  • $\begingroup$ Looks good to me, except I think you mean $g \in B(f, \epsilon)$, not $0.9\epsilon$. $\endgroup$ – Nate Eldredge May 26 '12 at 1:30
  • $\begingroup$ @NateEldredge I think that you mean $g\in \overline{B(f,\epsilon)}$ not $g\in B(f,\epsilon)$. $\endgroup$ – Gaston Burrull May 26 '12 at 1:32
  • $\begingroup$ Yes, you're right. $\endgroup$ – Nate Eldredge May 26 '12 at 1:33
  • $\begingroup$ Oops I think that may be $1.1\epsilon$ indeed $0.9\epsilon$ $\endgroup$ – Gaston Burrull May 26 '12 at 1:37
3
$\begingroup$

Here’s a similar argument that I find a bit easier. Let $$f(x)=\begin{cases}\frac1{1+x^2},&\text{if }x\in\Bbb Q\\\\0,&\text{otherwise}\;.\end{cases}$$

For $a>0$ let $$f_a(x)=\begin{cases}f(x),&\text{if }|x|\le a\\\\0,&\text{otherwise}\;.\end{cases}$$

Clearly $f\in X$, each $f_a\in\mathcal{B}(\Bbb R,\Bbb R)\setminus X$ (since $f_a$ is continuous on $\Bbb R\setminus[-a,a]$), and

$$\|f-f_a\|=\frac1{1+a^2}$$ can be made arbitrarily small by taking $a$ large enough.

$\endgroup$
  • $\begingroup$ Really good example! is less artificial that taking $f(x)=+x$, and $f(x)=-x$ and forcing boundness by more cases. I liked. $\endgroup$ – Gaston Burrull May 26 '12 at 3:12

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.