Discussion between Mary Star and Thomas

StackExchange

16:07

You defined a map from $\mathbb{R}\times S^1$ onto the catenoid. As a parametrization this is fine. When searching for an atlas then the problem with this is that the source is not diffemorphic to Euclidean space but to a cylinder. Just cover the $S^1$ in your domain of definition by two line se...

Mary Star

Why isn't the source diffemorphic to Euclidean space but to a cylinder?

Thomas

@MaryStar The circle $S^1$ is not diffeomorphic to $\mathbb{R}$, but it contains closed loops which cannot be deformed smoothly into a single point. Since it enters a factor of $\mathbb{R}\times S^1$ the product cannot be diffeomorphic to $\mathbb{R}^2$ either. Strictly speaking you don't need to know that for your task, but you have to write down the domain of definition for your charts. Even if you don't know that $\mathbb{R}\times S^1$ is not diffeomorphic to Euclidean space it is not clear that it is suitable as domain of definition for a chart. So you should try to find something else.

Mary Star

I got stuck right now... Could you explain to me how we search for an atlas?

Thomas

@MaryStar You have to look for coordinate patches which are homeomorphic to Euclidean space (e.g. Euclidean space itself, a unit disc, a product of intervals, ...). In the present case I already provided a possible choice, and the maps are also already given.

Mary Star

Why do we cover the $S^1$ in the domain of definition by these two line segments?

Thomas

16:07

@MaryStar Because (each of) these line segments is/are diffeomorphic to an interval, while $S^1$ is not.

Mary Star

Ahaa... Ok... $$$$ Consider we have the half-cone $$S: \ x^2+y^2=z^2, z>0$$ It can be parametrized by $$\sigma (u,v)=(u\cos v, u\sin v, u)$$ Do we take a subset of the domain of definition so that it is homeomorphic to a unit disc, to find the atlas in this case?

Thomas

@MaryStar Two subsets will do, one won't suffice in this case. You can also use more than two (as many as you want, actually), you only have to cover the manifold with the images of the maps, which should be defined on something diffeormorphic to Euclidean space (this is what is usually used in the definitions).

Mary Star

Which is the original domain of definition of the half-cone? $[0,2\pi]\times [0,2\pi]$ ?

Thomas

@MaryStar You have written down the parametriziation so you should decide, there is no general rule. I'd take $[0,2\pi)\times (0,\infty)$ (which is not nice since it has a boundary) or $(0,2\pi)\times (0,\infty)$ allowing the lack of one ray. If you take $(0,2\pi)\times (0,\infty)$ and $(-\pi,\pi)\times (0,\infty), $ say, you'd get your atlas.

Mary Star

I see... Thank you so much for your help!! :-)

Could I ask you also an other question?

I am looking at the following exercise:

Let $S$ be the half-cone $x^2+y^2 = z^2, z > 0$.
Define a map $f$ from the half-plane $\{(0, y, z)| y > 0\}$ to $S$ by $f(0, y, z) = (y \cos z,y \sin z,y)$.
Show that $f$ is not a diffeomorphism.

I have done the following:

To see that $f$ is a local diffeomorphism, we parametrize the half-plane $\{(0, y, z)| y > 0\}$ by the surface patch $\mathcal{\pi}(u, v) = (0, u, v)$, and use the atlas $\{\sigma|_{U_1} ,\sigma|_{U_2} \}$ where $$U_1=\{(u,v)\in \mathbb{R}^2 \mid u>0, 0<v<2\pi\} , \ \ U_2=\{(u,v)\in \mathbb{R}^2\mid u>0, -\pi<v\pi\}, \\ \sigma(u…

Thomas

16:39

This is hard to read here in the comments section. The set $f(\mathcal{O})$ with your choices is the cone minus a ray on the cone which is perpendicular to the circles which make up the cone. The map $f$ is not a diffeomorphism since it covers the cone several times.

Mary Star

What is a "ray"? @Thomas

With that what I have done we conclude that f is a local diffeomorphism, right?

4 hours later…

Thomas

20:27

A ray is half of a straight line. Yes, with what you've done you can conclude it's a local diffeomorphism.

Mary Star

Ah ok... And which is the equation of that ray? Is this ray on the $z$-axis? @Thomas

last day (15 days later) »

all rooms