qedbot

Wikipedia·wikipedia:MovingSofa

Moving Sofa Problem

conjecture formal record: mixed F2 declared

Machine-checked by Palomar.

Fidelity F2: The correspondence is declared through a Comparator challenge, an alignment table and written divergences.

Source

AI activity

How grades work

No AI contribution recorded against this statement.

F2 declared. The correspondence is declared through a Comparator challenge, an alignment table and written divergences.

declares divergences from its source reviewed by an agent

Checks

1
  • verified·Palomar

    Registered by Palomar at 1174c5c2: Comparator confirmed 17 theorems prove the recorded statement within Palomar's axiom policy, replayed through Lean's kernel and the independent nanoda kernel

    project registered at a pinned commit·2026-10-08·commit 1174c5c25c1c

    PALOMAR-2026-10-02-000008

Declared by the projects

3

As each project's formalization.yaml states it.

Optimality of Gerver's sofadeancureton/MovingSofa · joined by artifact · no independent check

Read formalization.yaml

authors
Dean Cureton
method
agent
review
agent-reviewed (Codex and Claude subagent reviewers (mathematical review of each proof, and statement checks against the paper), Claude subagent auditors (final audit, statement audit, whole-paper scope audit))
axioms
Classical.choice, Quot.sound, propext
sorry
0 unproved goals declared
sources
Optimality of Gerver's Sofa — formalizes, authors not-contacted; On moving a sofa around a corner — adapts, authors not-contacted; Differential equations and exact solutions in the moving sofa problem — background, authors not-contacted
divergences
Compared with formal-conjectures: MovingSofaSubmission/Challenge.lean copies the formal-conjectures definitions and both statements unchanged and imports only Mathlib. It inlines the plane notation and instances, names the rigid-motion topology instance so that Comparator sees the same name in Challenge and Solution, and leaves out the tests, metadata and open shape-uniqueness conjecture found there. Compared with the paper: an AI review of the statement found the following differences, none of which weakens the theorem. Sofas satisfy m 0 = id and start inside the horizontal arm, where the paper allows any initial translation; canonical_paper_motion_bridge proves that a sofa of either kind is a translate of one of the other kind. Sofas are closed and connected, as in the paper's Definition 1.2, so the theorem says nothing about disconnected sets. Motions lie in the full isometry group E(2), where the paper uses SE(2); orientation preservation is proved. Boundedness and measurability are not assumed; every moving sofa is proved compact. The supremum is taken in the extended nonnegative reals and is proved finite, at least 11/5 and attained. ABφθSpec uses non-strict inequalities, whi
checked by
nobody independent of its authors yet
Optimality (Baek 2024) and uniqueness of Gerver's sofa, formalized in Lean 4vltanh/lean4-moving-sofa/baek · joined by anchor · no independent check

Read formalization.yaml

authors
The-Anh Vu-Le
method
agent — ChatGPT Pro 6 (OpenAI), Claude Opus 5.5 (claude-opus-5-5)
review
agent-reviewed (Claude Opus 5.5 sub-agents, separate from the proving agents: a statement-fidelity review against the LaTeX source, and an audit of the paper's findings against the LaTeX source, A Claude Opus 5.5 sub-agent, separate from the proving agents: a review of the uniqueness statements against the informal proof)
sorry
0 unproved goals declared
sources
Optimality of Gerver's Sofa — formalizes, authors not-contacted; Uniqueness of Gerver's sofa: complete paper argument — formalizes, authors n/a; Differential Equations and Exact Solutions in the Moving Sofa Problem — background, authors not-contacted; On Moving a Sofa around a Corner — background, authors not-contacted
divergences
(1) Corrections of the paper's statements, each the minimal change that makes the statement true and each explained in the docstring of the corresponding declaration: Definition 2.3.9 ("x-axis" for "y-axis"); Proposition 2.2.2 (the missing "-1" in Q^-); Proposition 2.5.4 (the exchanges a<->c, b<->d, W<->Z, and K for K^m on the right); Definition 3.2.5 (the fan F_omega for P_omega); Proposition 3.3.1 (2) (the bottom normals omega + pi, 3 pi/2); Lemma 4.2.4 (omega in [arcsec 2.2, pi/2)); Proposition 5.1.4 (integrable instead of bounded density; proposition5_1_4_as_stated_false refutes the printed version); Proposition 6.2.2 (argument pi/2 - t); Lemma 6.5.5 (range (0, pi/2]); Lemma 8.1.7 (4) (phi^L for phi^R); Definition 8.3.3 (pi + phi^R for pi/2 + phi^R); Lemma 8.3.6 (3) (evaluation at pi/2); Theorem 8.4.3 (2) (B(t3) for D(t3)); Proposition 8.4.4 (4) (shifted by pi/2); Theorem 3.1.2 (the Nef polygon is bounded, a hypothesis that the paper leaves out; baek/REPORT.md, Section 4); Lemma 7.1.6 (in a real vector space, where K + c is defined, for every convex-bilinear h; E18). (2) Readings: the plane is R x R and area is Lebesgue measure; a rigid motion is a continuous angle and translat
checked by
nobody independent of its authors yet
Optimality, uniqueness and stability of Gerver's sofa from one coercive certificate, formalized in Lean 4vltanh/lean4-moving-sofa · joined by anchor · Palomar

Read formalization.yaml

authors
The-Anh Vu-Le
method
agent — ChatGPT Pro 6 (OpenAI), Claude Opus 5.5 (claude-opus-5-5)
review
agent-reviewed (A Claude Opus 5.5 sub-agent, separate from the repairing agents: a comparison of every repaired stability proof with the uncompiled draft's, Claude Opus 5.5 sub-agents, separate from the writing agents: comparisons of the manuscript's Sections 10 and 11 with the Lean, A Claude Opus 5.5 sub-agent that could not edit: a comparison of the Challenge's certificate definitions with Baek's paper and the libraries)
sorry
0 unproved goals declared
sources
Optimality of Gerver's Sofa — adapts, authors not-contacted; Stability of Gerver's sofa: notes — formalizes, authors n/a; The coercive route: notes — formalizes, authors n/a; Uniqueness of Gerver's sofa: complete paper argument — adapts, authors n/a
divergences
(1) The Challenge's definitions of Q and of the enlarged domain follow Baek's paper as the repository formalizes it: Q is Baek's Definition 8.2.2, the curve areas are Baek's Definitions 7.2.6 and 7.2.8, the convex curve areas his Theorem 7.3.2, the caps his Definition 2.4.1, and the enlarged domain is Baek's domain of triples with the cap in the set of all caps with rotation angle pi/2 instead of the caps with the injectivity condition (the stability notes' definition). (2) Readings, as in versions 1 to 4: the plane is R x R and area is Lebesgue measure; the support function is a supremum; the surface area measure sigma_K is the Lebesgue-Stieltjes measure of t -> <v_K^+(t), v_t> + int_0^t h_K; the curve area functional of a continuous curve of bounded variation is (1/2) int x(t) x dx(t), with dx the Lebesgue-Stieltjes vector measure of the curve on [a, b]; Gerver's cap is the cap of Gerver's sofa, P_{pi/2} intersected with the closed quarter-planes of its supporting hallways. In the Challenge, sigma_K is defined by a case split on the monotonicity and right continuity of its distribution function, since a Challenge cannot carry the library's proofs of these properties; the Solution
checked by
Palomar

Follow and discuss

All discussion

Discussion and bounties for this problem load here.

Something wrong or missing here? Request a correction or add a claim, with its sources.

Formal material

Formal statements · 1
Cited proofs · 13
Also known as · 2
  • https://github.com/google-deepmind/formal-conjectures/blob/main/FormalConjectures/Wikipedia/MovingSofa.lean
  • FormalConjectures/Wikipedia/MovingSofa.lean

Cite this record

qed.bot, “Moving Sofa Problem”, https://qed.bot/s/wikipedia-movingsofa, as of 20 Nov 2026.

This record as plain text, with each claim, its grades and its sources.