Zulip Chat Archive

Stream: general

Topic: Discussion: Carleson project


Floris van Doorn (Sep 01 2025 at 15:48):

This is a discussion thread for the Carleson project for everyone not in the #Carleson channel.

Floris van Doorn (Sep 01 2025 at 15:53):

The announcement is here: #announce > Carleson project

Floris van Doorn (Sep 01 2025 at 15:54):

An article about this project also appeared in the German popular science magazine Spektrum: https://www.spektrum.de/news/so-veraendern-computer-ki-und-beweispruefer-die-mathematik/2280992 (unfortunately behind a paywall)

Henrik Böving (Sep 01 2025 at 16:00):

Neben den ZFC-Axiomen gibt es auch andere Grundregeln, die als Fundament der Mathematik dienen können. Das Computerprogramm Lean fußt zum Beispiel auf der Homotopic Type Theory

They wrongly claim we are HoTT based

Floris van Doorn (Sep 01 2025 at 16:01):

Yeah, I also noticed that error. I did mention during the interview that the foundation of Lean was similar to HoTT, which the interviewer had already heard about. So I guess that's how it slipped in.

Ruben Van de Velde (Sep 01 2025 at 16:15):

How much are you planning to upstream?

Floris van Doorn (Sep 01 2025 at 16:20):

The ToMathlib folder contains ~10k (out of ~35k) lines of code, so approximately that much.
Upstreaming discussion is happening here: #Carleson > Porting to Mathlib

Jakob von Raumer (Sep 01 2025 at 16:56):

Floris van Doorn said:

An article about this project also appeared in the German popular science magazine Spektrum: https://www.spektrum.de/news/so-veraendern-computer-ki-und-beweispruefer-die-mathematik/2280992 (unfortunately behind a paywall)

For ze Germans who don't remember how to tear down walls: https://archive.is/AkjaJ

Filippo A. E. Nuccio (Sep 01 2025 at 21:35):

Jakob von Raumer said:

Floris van Doorn said:

An article about this project also appeared in the German popular science magazine Spektrum: https://www.spektrum.de/news/so-veraendern-computer-ki-und-beweispruefer-die-mathematik/2280992 (unfortunately behind a paywall)

For ze Germans who don't remember how to tear down walls: https://archive.is/AkjaJ

Danke!

Jeremy Tan (Sep 02 2025 at 01:31):

@Floris van Doorn I think you promised to include all of the authors of the blueprint paper in the other paper? (The other authors, including me, are not listed on https://arxiv.org/pdf/2508.05563)

Kim Morrison (Sep 02 2025 at 06:58):

Congratulations. :-)

So... who is going to start a project that needs the main theorem as a prerequisite? I want to see Mathlib grandchildren in action! :-)

Floris van Doorn (Sep 02 2025 at 10:49):

Jeremy Tan said:

Floris van Doorn I think you promised to include all of the authors of the blueprint paper in the other paper? (The other authors, including me, are not listed on https://arxiv.org/pdf/2508.05563)

This was never the plan. The plan is to write a paper specifically about the formalization, I'll start coordinating that in a couple of weeks. That paper will have as authors the authors of the formalization.
This paper is about the mathematical result, and only contains the fact that it is formalized as a remark.
I hope that clears things up.

Jeremy Tan (Sep 03 2025 at 07:29):

:squared_ok:


Last updated: Dec 20 2025 at 21:32 UTC