Zulip Chat Archive

Stream: maths

Topic: Tietze extension theorem


Yury G. Kudryashov (Nov 04 2021 at 00:57):

I'm formalizing Tietze extension theorem. Most sources state if for f : s → ℝ. Clearly, it is true for f : s → (ι → ℝ) and f : s → E, where E is a finite dimensional normed space over . Are there more general versions? E.g., is it true for f : s → E, where E is a Banach space?

Scott Morrison (Nov 04 2021 at 01:14):

https://mathoverflow.net/questions/201704/between-tietzes-and-dugundjis-extension-theorems

Scott Morrison (Nov 04 2021 at 01:16):

Oh, maybe that's not quite it. I thought Bill Johnson's answer was addressing your question, but now I think the hypotheses may be different.

Yury G. Kudryashov (Nov 04 2021 at 01:19):

BTW, it seems that the proof at ncatlab in the case of an unbounded function has a mistake (or I miss something trivial).

Finally consider the case that ff is not a bounded function. In this case consider any homeomorphism ϕ ⁣:R1(c0,c0)R1\phi \colon \mathbb{R}^1 \overset{\simeq}{\to} (-c_0,c_0) \subset \mathbb{R}^1 between the real line and an open interval. Then ϕf\phi \circ f is a continous function bounded by c0c_0 and hence the above argument gives an extension ϕf^\widehat {\phi \circ f}. Then ϕ1ϕf^\phi^{-1} \circ \widehat{ \phi \circ f } is an extension of ff.

Why ϕf^\widehat {\phi \circ f} can't take value c0c_0 at some point?

Scott Morrison (Nov 04 2021 at 01:21):

https://mathoverflow.net/questions/136554/generalizations-of-the-tietze-extension-theorem-and-lusins-theorem?rq=1 seems to be about your question.


Last updated: Dec 20 2023 at 11:08 UTC