0

Problem from Conway's "A course in point set topology": If for $k=1,2, A_k$ is a dense subset of the metric space $(X_k,d_k)$, then show that $A_1\times A_2$ is a dense subset of $X_1\times X_2$.

My attempted solution: Suppose $A_k\subseteq X_k$ is dense $\Rightarrow cl(A_k)=X_k\Rightarrow \forall r_k,\space B(x_k,r_k)\cap A_k\neq \varnothing$ so take $y_k\in B(x_k,r_k)\cap A_k$. Take $r=max\{r_1,r_2\}$ so that for $(y_1,y_2)\in B((x_1,x_2),r)\Rightarrow B((x_1,x_2),r)\cap A_1\times A_2\neq \varnothing$ whenever $(y_1,y_2)\in B((x_1,x_2),r)\cap A_1\times A_2\Rightarrow (y_1,y_2)\in cl(A_1)\times cl(A_2)\Rightarrow (y_1,y_2)\in X_1\times X_2$ since both $A_1$ & $A_2$ are both dense $\Rightarrow cl(A_1\times A_2)=X_1\times X_2$

Please drop a line if you think my proof is right, and of course, let me know how to improve

Jaider
  • 413
  • 3
    Overuse of the $\Rightarrow$ symbol hurts readability. I suggest not using it at all. Use words and normal English sentence structure instead. – Nate Eldredge Jun 14 '20 at 19:57
  • You could use the fact that the closure of the product is the product of the closures – Someone Jun 14 '20 at 19:59
  • I know, its not my final write up. Just trying to get the ideas flowing together – Jaider Jun 14 '20 at 19:59
  • I'm going off of what I learned up to this point in the book to prove this – Jaider Jun 14 '20 at 20:01
  • 3
    But it's hard for someone to offer comments on your solution if they can't follow it. – Nate Eldredge Jun 14 '20 at 20:01
  • Oh ok , I have not read those notes on topology, but I would recommend Munkre's book – Someone Jun 14 '20 at 20:02
  • 1
    Your use of $\Rightarrow$ is not logically correct. If you must use symbols rather than words, then the symbol you want is $\therefore$. $\phi \Rightarrow \psi$ means that if $\phi$ is true then so is $\psi$: it is neutral about the actual truth value of $\phi$. – Rob Arthan Jun 14 '20 at 20:13
  • I see, thanks for letting me know. I'm still pretty raw and tend to overuse the amount of math symbols. – Jaider Jun 14 '20 at 20:17
  • What metric are you using on the product? The max-one ? Then you need $\min {r_1,r_2}$ not the max..draw a picture.. – Henno Brandsma Jun 14 '20 at 22:12
  • Can't remember off the top of my head, drawing a picture first probably would've helped, but I think I used it to make sure it encompassed both balls or something along that idea – Jaider Jun 14 '20 at 22:26
  • But would that really matter anyway? I tried drawing a diagram and it seems that either the minimum or maximum would both work just the same – Jaider Jun 14 '20 at 23:03

1 Answers1

3

I’m afraid that your argument is so badly organized and confusingly written that I can’t tell whether it’s correct. From its conclusion it appears that you’re trying to prove that $\operatorname{cl}(A_1\times A_2)=X_1\times X_2$. You should begin by saying so, since there are other ways to prove that $A_1\times A_2$ is dense in $X_1\times X_2$.

To prove that $\operatorname{cl}(A_1\times A_2)=X_1\times X_2$, you could either prove the general result that $\operatorname{cl}(H\times K)=(\operatorname{cl}H)\times\operatorname{cl}K$ for arbitrary $H\subseteq X_1$ and $K\subseteq X_2$ or prove this special case. To do the latter you should start with an arbitrary point $\langle x_1,x_2\rangle\in X_1\times X_2$ and show that every open nbhd of this point intersects $A_1\times A_2$, something like this:

Let $p=\langle x_1,x_2\rangle\in X_1\times X_2$, and let $U$ be any open nbhd of $p$. Then there are real numbers $r_1,r_2>0$ such that $p\in B(x_1,r_1)\times B(x_2,r_2)\subseteq U$. $A_1$ is dense in $X_1$, so there is a $y_1\in B(x_1,r_1)\cap A_1$; similarly, there is a $y_2\in B(x_2,r_2)\cap A_2$. Clearly $$\langle y_1,y_2\rangle\in \big(B(x_1,r_1)\times B(x_2,r_2)\big)\cap(A_1\times A_2)\subseteq U\cap(A_1\times A_2)\;.$$ Thus, every open nbhd of $p$ contains a point of $A_1\times A_2$, so $p\in\operatorname{cl}(A_1\times A_2)$. And $p$ was an arbitrary point of $X_1\times X_2$ so $\operatorname{cl}(A_1\times A_2)=X_1\times X_2$.

Note that I’ve used a lot more words than you did: I use them as connective tissue to make the flow of logic clearer. I agree completely with Nate Eldredge’s comments: except in formal logic a proof should be written in ordinary prose, with symbols only where they are necessary or genuinely make things clearer. (For instance, the quadratic formula is much clearer when written in algebraic symbols than when it is expressed in words!)

Brian M. Scott
  • 631,399
  • Hi professor Scott, could I ask your assistance here? I think I found a good answer but I'm not sure about. – Antonio Maria Di Mauro Jun 14 '20 at 21:12
  • Thank you for the help on the problem, as well as for the advice, I just find it natural to condense an answer down as simple as possible without using much language to get your point across. – Jaider Jun 15 '20 at 00:08
  • @Jaider: You’re welcome. The problem with that is that it makes it hard for others to read; of course this doesn’t matter in one’s notes for one’s own use, but it’s a bit of a problem when one is writing things up for others to read. – Brian M. Scott Jun 15 '20 at 02:13