Homeomorphism

anyone cares to explain why these two are not topologically equivalent? The second figure does not have the 'tip' included.
they both have the infinite number of 2-points (the vertical line) and no 3-points.

It seems that if you shrink the vertical line down to a point, then the figure on the right will have a hole but the one on the left won't. I can't remember if this is homeomorphism or some homotopy condition.

Is this the tip just one point? Like the left one is a circle with a half-open interval attachhed and the second one is with a closed interval attached?

If the tip is like a representation of a closed interval like says then the one in the left isn't compact but the ritght one is

yeah, that was what I was gonna say and it is the right answer and it is trivial and OP should have thought of that. No offense, it's just sometimes I think about using big fancy theorems while there's an easy solution right in front of me.

The sets have the subset topology of [math]\mathbb{R}^2[/math]. If two set are homeomorphic and one is compact, then the other is compact as well. As already noted, one of the two sets is not compact, while the other is.
1. One is closed (prove this) and limited hence compact (Heine-Borel).
2. The other is not compact: use the same reasoning you use for [math]\{1/n\}[/math].

what is the quickest way of proving that the left one is closed ?

Prove that the circle is closed (easy, since preimage of a closed point under a continuous map) and that an interval embedded in R^2 is closed. Then the left one is a union of two closed sets and hence closed.

>why is an open interval not homeomorphic to a closed interval

rly mak eu think

>what is the quickest way of proving that the left one is closed ?
Because it's equal to it's closure (show this*), which is an intersection of closed sets, hence closed.
Hope I'm not mistaken.
* See the definition of closure and use circles as open sets.
In the same manner you can check whether the other set is open (using the interior instead).

>it's
its.

>>why is an open interval not homeomorphic to a closed interval
Let [math]X[/math] and [math]A,B\subset X[/math] with the subset topology.
[math]A[/math] is open and [math]B[/math] is closed.
[If they are bounded, [math]A[/math] cannot be compact (at least in this case for Heine-Borel) thus the sets are not homeomorphic.]
What is the general proof?

dude what ?

>dude what ?
Did I make a mistake or the post is not clear?

>what is the general proof?
general proof of what please?

>general proof of what please?
Is this
>Let X and A,B⊂X with the subset topology.
>A is open and B is closed.
>A and B cannot be homeomorphic.
true?

consider A=B=X

It's definitely not true: let A = B = {empty set}. I think it might be true if you insist that X is connected and A and B are proper nonempty subsets, but I can't think of a proof or counterexample at the moment.

Thank you.
So, in this case, are there simpler alternatives to:
1. showing that one of the sets is open, one is closed and invoking Heine-Borel,
2. showing by hand that one is compact and the other is not?

In this particular case, you may notice that removing any point from an open interval makes it disconnected, whereas in a closed interval, you can remove the boundary points and it will stay connected.

no, but it's simple enough. it really boils down to showing that the left subset is not closed while the second one is. the spaces are so nicely embedded that this can be proven straight from the definition using euclidean or equivalent metric, you won't even get your hands dirty.

Go it, but one last question.
When can subsets of [math]\mathbb{R}^{2}[/math] be called intervals?
Something like being homeomorphic to intervals in [math]\mathbb{R}[/math]?

was reacting to

well either you consider two-dimensional intervals which are just rectangles but this is obviously not the case.

the correct term would be "embedded interval" which is an image of an embedding of some interval. this is of course equivalent to being homeomorphic to an interval in the subspace topology.

I don't think it's really common terminology. That being said, you can define a segment in the plane as the convex hull of two points, ie. [math][A,B] = \{tA + (1-t)B, t \in [0,1]\}[/math]

I'm asking because you said
>that removing any point from an open interval...
and I was wondering by what definition are the sets in OP's post considered intervals.
I otherwise completely understand you post referred to the real line.

Oh, yeah I did mean in the real line. In OP's pic, I think the compactness argument is the best one (which is not to be phrased as "uhh one of them is closed and the other is not" because being "closed" is a relative property, whereas compactness is an intrinsic property)

Thank you both and thanks to OP.

compactness makes sense. what about these two? compactness clearly doesn't work in this case

there should exist an argument using just connectedness but I don't see it right now. anyway, those spaces have different fundamental groups. do you know what that is?

does the fact that one set of 2-points is closed and the other is not? i.e. we can't have a homeomorphism between those sets, and if we had a homeomorphism between those two figures and reduce it to 2-points, it'd still have to be homeomorphism

ugh, I think I haven't explained myself well here.

My reasoning is as follows:
suppose we have a homeomorphism between the two figures. call it f. we know that if x is a 2-cut point in figure A, f(x) is a 2-cut point in figure B. so by reducing f to set of 2-cut points in figure A, we see that there can't exist a homeomorphism as one of the sets is closed and the other is not

now I know, that works. thanks!

got something. on the bottom one, there is a point with the following property: every connected neighborhood of the point remains connected after the removal of the point. the upper one doesn't have such point.

Second figure is compact.

Eh, I couldn't read your drawings. First one is compact.