# Uniqueness of smooth structure on a zero-dimensional smooth manifold

In John Lee's Book "Introduction to Smooth Manifolds" on page 17, Example 1.12, the author states that the smooth structure on any zero - dimensional manifold is unique. That confuses me, suppose for example the $M = \{ p \}$ is a point. Then I can for example give this manifold a smooth structure by taking $(U, \psi)$ to be given by $$U = \{ p \}, \quad \psi(p) = 1$$ or, I could also give it the structure $$U = \{ p \}, \quad \psi(p) = 2$$ From what I understand, these are different smooth structures. What am I missing ?

-
$\mathbf R^0$ is just a point, isn't it? Any transition maps are going to send that point to itself. – Dylan Moreland Apr 18 '12 at 20:18
@Dylan: Suppose M is the set $\{A,C,T,G\}$ the bases for DNA. I can identify this set with any discrete set consisting of four real numbers. What I don't understand is, to what extend is any such identification unique ? – harlekin Apr 18 '12 at 20:57
You're off by one dimension. the charts should map to $R^0$, not $R^1$. – David Dec 13 '15 at 15:31

As Dyland Moreland points out $\mathbb{R}^0=\{p\}$. Thus, there is for each discrete space $X$ a unique zero-dimensional manifold structure with the charts $\psi_x:\{x\}\to\mathbb{R}^0$ being the unique such maps.
@Youcis: In what sense do you mean $\mathbb{R}^0 = \{p\}$ ? I thought that $R^0$ stands for any discrete set, not necessarily a one - point set ? – harlekin Apr 18 '12 at 21:01
Well, then the symbol $\mathbb{R}^0$ wouldn't make any sense. No, $\mathbb{R}^0$ is the $0^{\text{th}}$-dimensional Euclidean space which is just a point. – Alex Youcis Apr 18 '12 at 21:04
@harlekin If you mean how does it allow you to locally think about Euclidean space that's simple. To think about something locally being $\mathbb{R}^0$ means it locally looks like a point, which means that it's discrete. If you meant coordinate in the sense that you can do interesting things, like calculus, you really can't. While you can define a zero-dimensional manifold $\mathbb{R}^0$ doesn't carry nearly enough structure to do complicated things like calculus. Does that help? – Alex Youcis Apr 18 '12 at 21:12
@Youcis: hm .. ok - I need to think about this, your comment was very helpful! To help clarify things - I know there are standard structures on Euclidean Spaces $\mathbb{R}^n$, given by the atlas consisting of one element, namely $(U,\psi) = (\mathbb{R}^n, \text{Id})$ (where Id stands for the identity map). So, for $n = 0$, what is the standard coordinate that $p$ gets send to ? Is that $0$ ? Sorry for asking such a stupid question - I realize that I don't really understand $\mathbb{R}^0$ - thanks for your patience ! – harlekin Apr 18 '12 at 21:22