3 ms·
Oh cool, thanks! > - Any subset which does not contain infinity We get those from the original (discrete) topology, I assume? > - Any subset containing infin
by Twisol 5y ago
Oh cool, thanks!
> - Any subset which does not contain infinity
We get those from the original (discrete) topology, I assume?
> - Any subset containing infinity whose complement is finite.
Ohh, is this construction related to the cofinite topology, then? I always wondered what that was useful for. (I never quite made it all the way through Munkres.)
Also, I start to see why Wikipedia characterizes this as `{0} u {1/n | n in N}`. Any ball containing 0 will miss only finitely many points.
> So, in order for an endomap to be continuous, [...]
It took me a bit to work through this from first principles, but I see! Thanks for lighting the path.
- kevinventullo 5y ago> We get those from the original (discrete) topology, I assume? Yep! > ... is this construction related to the cofinite topology, then? I’m not sure, it’s sort of only cofinite “at infinity”? One place you see the cofinite topology appear is in algebraic geometry. The Zariski topology of a (connected) one-dimensional algebraic variety is the same as the cofinite topology. Glad I could help!