Zulip Chat Archive

Stream: lean4

Topic: What fragment of Lean is at least weakly normalizing?


nrs (Oct 09 2024 at 02:19):

Am interested in using Lean as a correct-by-construction general programming language, and I've read in the article about Lean4Lean that it fails to be strongly normalizing. Do we know what fragments are at least weakly normalizing?

Mario Carneiro (Oct 11 2024 at 17:09):

if you exclude Acc and never reduce proofs it's plausible that the resulting system is strongly normalizing

Patrick Massot (Oct 11 2024 at 19:42):

I’m not an expert in those question but I guess the two parts of your first sentence are completely unrelated. The strong normalization property does not prevent Lean from doing correct by construction programming.


Last updated: May 02 2025 at 03:31 UTC