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

Talking about an example of an open dense subset of $[0,1]$ with measure $1/2$ I heard talking about 'fat rationals' but I don't find anything on the internt.. Does someone know about this?

share|cite|improve this question
Hmmm... Never heard "fat rationals", though I've heard "fat Cantor set". You can take a fat Cantor set of measure $1/2$; it's complement will be open. The fat Cantor set has empty interior, so its complement is dense. – Arturo Magidin Apr 17 '12 at 20:12
'fat rationals' probably means a set of the form $$A = \bigcup_{n\in \mathbb N} B_{\epsilon_n}(q_n)$$ where $\{q_n\}_{n\in \mathbb N}$ is an enumeration of the rationals and the $\epsilon_n>0$ are chosen in such a way that the measure of $A$ is $1/2$. – Sam Apr 17 '12 at 20:21
up vote 4 down vote accepted

For a fixed $\varepsilon$, consider the set $S_{\varepsilon}:=\bigcup_{n\in\mathbb N}(q_n-2^{-n}\varepsilon,q_n+\varepsilon 2^{-n})$, where $\{q_n\}$ is an enumeration of the rationals. Then $S_{\varepsilon}$ is open and dense in $[0,1]$, since it contains all the rationals of this interval. The maps $f\colon\varepsilon\mapsto \lambda(S_{\varepsilon})$ is Lipschitz-continuous. Indeed, if $\varepsilon_1\leq\varepsilon_2$, we have \begin{align*}f(\varepsilon_2)-f(\varepsilon_1)&=\lambda(S_{\varepsilon_2}\setminus S_{\varepsilon_1})\\\ &\leq \lambda\left(\bigcup_{n=0}^{+\infty}(q_n-2^{-n}\varepsilon_2,q_n+\varepsilon_2 2^{-n})\setminus (q_n-2^{-n}\varepsilon_1,q_n+\varepsilon_1 2^{-n})\right)\\\ &\leq \sum_{n=0}^{+\infty}\lambda((q_n-2^{-n}\varepsilon_2,q_n+\varepsilon_2 2^{-n})\setminus (q_n-2^{-n}\varepsilon_1,q_n+\varepsilon_1 2^{-n}))\\\ &=2(\varepsilon_2-\varepsilon_1)\sum_{n=0}^{+\infty}2^{-n} \end{align*}

share|cite|improve this answer
The continuity consideration of $\varepsilon \mapsto \lambda(S_\varepsilon)$ implies of course that by suitably choosing $\varepsilon$ you can obtain any desired real value $\gt 0$ as measure of $\lambda(S_\varepsilon)$. – t.b. Apr 17 '12 at 21:09

As an alternative to Davide Giraudo's proposal which avoids having to order the rationals, try

$$S_k = [0,1] \cap \bigcup_{a\in\mathbb Z, b\in \mathbb Z^+}\left(\frac{a}{b}-\frac{k}{b 2^{b}},\frac{a}{b}+\frac{k}{b 2^{b}}\right).$$

For $k\approx 0.32184058\ldots$ this will have measure $\frac12$.

This has the same continuity considerations as Davide's and I suspect that as a function of $k$ the measure of $S_k$ has a zero derivative for all rational $k$, though I have not checked.

share|cite|improve this answer
How did you think of this one?! And what about the value of k? – balestrav Apr 18 '12 at 23:08
@balestrav: It is similar to something I did earlier with dyadic fractions. Working out the measure is simply a question of looking at some overlapping intervals (I took $b$ up to $50$), and then iterating until the desired measure target is achieved. It is nice that the intersection of any interval $[x,x+1]$ with the union produces the same measure. – Henry Apr 18 '12 at 23:53
That page has now moved to – Henry Mar 2 at 19:54

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.