Is the following argument correct?

Let $D_1$ and $D_2$ be any open discs in $\mathbf{R}^2$ with $D_1\cap D_2\neq\varnothing$. If $(a,b)$ is any point in $D_1\cap D_2$, show that ther exists an open disc $D_{(a,b)}$ with center $(a,b)$ such that $D_{(a,b)}\subset D_1\cap D_2$.

Proof. Let $D_1,D_2$ denote two arbitrary discs in $\mathbf{R}^2$ having $\alpha = (a_1,b_1)$ and $\beta = (a_2,b_2)$ as there centers and radii $r_1$ and $r_2$ respectively, that is \begin{align*} D_1 = \{(x,y)\in\mathbf{R}^2:(x-a_1)^2+(y-b_1)^2<r_1^2\}\\ D_2 = \{(x,y)\in\mathbf{R}^2:(x-a_2)^2+(y-b_2)^2<r_2^2\} \end{align*} Let $\gamma = (a,b)\in D_1\cap D_2$. We define the disc $D_{(a,b)}$ similar to $D_1,D_2$ but having center $(a,b)$ and radius $r = \min\{\frac{r_1-d_1}{8},\frac{r_2-d_2}{8}\}$, where $d_1$ and $d_2$ are defined as follows \begin{align*} d(\gamma,(\alpha,\gamma)) = d_1 = \sqrt{(a-a_1)^2+(b-b_1)^2}\\ d(\gamma,(\beta,\gamma)) = d_2 = \sqrt{(a-a_2)^2+(b-b_2)^2} \end{align*} Now let $(x,y)$ be an arbitrary point inside the disc $D_{(a,b)}$, appealing to the triangle inequality then yields \begin{align*} \sqrt{(x-a_1)^2+(y-b_1)^2} &=d(\alpha,(x,y))\leq d(\alpha,\gamma)+d(\gamma,(x,y))\\ &=d_1+r<d_1+\frac{r_1-d_1}{8}=\frac{r_1}{8}<r_1 \end{align*} \begin{align*} \sqrt{(x-a_2)^2+(y-b_2)^2} &= d(\beta,(x,y))\leq d(\beta,\gamma)+d(\gamma,(x,y))\\ &=d_2+r<d_2+\frac{r_2-d_2}{8}=\frac{r_2}{8}<r_2 \end{align*} Since our choice of $(x,y)$ was arbitrary, we have $D_{(a,b)}\subset D_1\cap D_2$.


  • 1
    $\begingroup$ The task is much easier if you prove it topologically. An intersection of two open sets is open. Since open balls form a base in $\mathbb{R}^2$, there will be an open ball inside that intersection. Although, you have to know the fact that open balls form a base in advance. $\endgroup$
    – Zeekless
    Oct 15 '18 at 18:21
  • 1
    $\begingroup$ It is correct for me $\endgroup$ Oct 15 '18 at 18:22
  • $\begingroup$ @Zeekless I am just starting topology, so would you care to elaborate what you mean by prove "topologically". $\endgroup$ Oct 15 '18 at 18:22
  • $\begingroup$ @Zeekless Thanks. $\endgroup$ Oct 15 '18 at 18:28
  • 2
    $\begingroup$ @Zeekless I think if you want to prove that the open balls form a base then you need to prove the statement from Atif Farooq (of course without using the fact that open balls form a base). So I dont think there is a way around doing the calculations that he did. $\endgroup$
    – supinf
    Oct 15 '18 at 18:36

The proof is correct and mostly good, there are only minor mistakes and typos.

You write $$ d_1+r<d_1+\frac{r_1-d_1}{8}=\frac{r_1}{8}<r_1 $$ and here the equality is false. But this can be easily fixed, for example (using $d_1< r_1$) $$ d_1+r<d_1+\frac{r_1-d_1}{8}=\frac{r_1}{4}<r_1 . $$ The same issue exist with $d_2,r_2$ below.

In my opinion it would be nicer if you choose $$ r= \min\{r_1-d_1,r_2-d_2\}, $$ then your proof would still work, but your choice is not wrong!

Small typos: You probably mean to write $d(\gamma,\alpha)$ or $d(\gamma,(a_1,b_1))$ instead of $d(\gamma,(\alpha,\gamma))$ and $d(\gamma,\beta)$ or $d(\gamma,(a_2,b_2))$ instead of $d(\gamma,(\beta,\gamma))$.

  • $\begingroup$ @Thank you so much for response, and i agree your choice of $r$ is much better than mine, aesthetics matter i guess. $\endgroup$ Oct 15 '18 at 18:38
  • $\begingroup$ Is it possible to prove it if the intersection is quadrilateral ambrsoft.com/TrigoCalc/Circles3/Intersection.htm (see case 12). Then for the other triangle,since Atif proved it for case 11 you would have to define d($a_3,b_3$)and d($a_4,a_4$) $\endgroup$
    – Eudoxus
    Dec 24 '20 at 13:06

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.