Zulip Chat Archive

Stream: maths

Topic: filter convergence and tendsto


Bernd Losert (Dec 14 2021 at 15:42):

How is filter convergence done in mathlib. For example, how do I say that the neighborhood filter at x converges to x? Do I write this as tendsto id (𝓝 x) (𝓟 {x}) ?

Johan Commelin (Dec 14 2021 at 15:53):

I think that "converging to x" typically means tendsto f _ (𝓝 x).

Johan Commelin (Dec 14 2021 at 15:53):

So in that case, "the neighbourhoud filter at x convergest to x" would be a tautology.

Bernd Losert (Dec 14 2021 at 16:08):

Ah, I think you're right. I had the filter arguments backwards. Do you know if there is special notation for this?

Bernd Losert (Dec 14 2021 at 16:09):

I was thinking of something like (𝓝 x) converges_to x.

Patrick Massot (Dec 14 2021 at 16:11):

No, there is no special notation.

Reid Barton (Dec 14 2021 at 16:12):

The notation is 𝓝 x ≤ 𝓝 x :upside_down:

Bernd Losert (Dec 14 2021 at 16:18):

I think you mean 𝓟 {x} ≤ 𝓝 x.

Patrick Massot (Dec 14 2021 at 16:20):

These are different statements.

Patrick Massot (Dec 14 2021 at 16:21):

The statement (𝓝 x) converges_to x is what Reid wrote.

Bernd Losert (Dec 14 2021 at 16:22):

OK. I guess I'm confused because I don't know what is playing the role of "x" here.

Bernd Losert (Dec 14 2021 at 16:23):

I was told that to think of filters as sets and ultrafilters as points, so 𝓟 {x} would be "x".

Bernd Losert (Dec 14 2021 at 16:25):

𝓟 {x} ≤ 𝓝 x <-- So this means that 𝓟 {x} converges to x.

Reid Barton (Dec 14 2021 at 16:25):

That's also true but it says that the principal filter at x converges to x--which probably is what you meant in the first place

Reid Barton (Dec 14 2021 at 16:28):

(As a sanity check, the notion "converges to x" depends on the topology, so it must involve 𝓝 x.)

Bernd Losert (Dec 14 2021 at 16:30):

Yeah. Thanks.


Last updated: Dec 20 2023 at 11:08 UTC