{"id":"0f525b73-67b8-4b2a-b20d-a3a9ffd9e602","arxiv_id":"2508.06741","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"New explicit upper bounds for Lp Poincare-Friedrichs constants of the de Rham complex over convex domains and over shellable triangulations, obtained from operator norms of regularized Poincare and Bogovski potentials.","lead":"These authors prove explicit, geometry-only upper bounds for the constants in Poincare-Friedrichs inequalities for gradient, curl, divergence, and general exterior-derivative problems, both on convex domains and on domains built from shellable triangulations. The result gives guaranteed lower bounds on vector-Laplacian and Maxwell eigenvalues without solving a global eigenvalue problem.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 8.2/8.3 use maximum height where only minimum is supported: the star-shaped ball can leave the domain, so the reflection step in Theorem 9.3 is not justified as written.","rationale":"The reader correctly identified Proposition 8.3 and its geometric prerequisites (Lemma 8.1, Proposition 8.2) as the most load-bearing part of the proof of Theorem 9.3. My stress-test agrees with that focus, but sharpens it: the actual defect is a min/max inconsistency inside Proposition 8.2 itself. The statement uses the maximum height of any vertex of S, while the proof uses the minimum height of z_S. For ℓ=0 these are different quantities, and the maximum can be arbitrarily larger than the minimum (e.g., a square with a deep notch gives max 10 and min about 2.1). Thus the stated ball may not even be contained in the star, let alone be a kernel ball with respect to which the star is star-shaped. Since Proposition 8.3 relies on precisely that ball to choose the point y and to ensure the reflected simplex remains inside U, the Jacobian bounds and the recursive estimate in Theorem 9.3 are not proven as written. I do not claim the main theorem is false; the construction can likely be repaired by using the minimum height and accepting larger constants. But as submitted, the proof of the central computable-bound claim has a concrete, testable gap. Therefore the appropriate verdict is CONDITIONAL: accept only after the geometric lemma is corrected and the constants in Section 9 are re-derived. This is not a rejection of the underlying approach, and it is not an ad hominem concern; it is a precise mathematical inconsistency in a key lemma.","tokens_in":57809,"tokens_out":23418,"duration_ms":252450,"concrete_test":"Implement the 2D fan triangulation of the notch domain: Ω = [-10,10]^2 minus the open triangle with vertices (1,10), (3,10), (2,1), with V=(0,0). Compute h as the maximum distance from V to any boundary edge of the star and compare B_h(V) with Ω. Also compute the distance from V to each boundary edge; if min ≈ 2.1 and max = 10, verify that B_10(V) contains points of the notch, contradicting Proposition 8.2. Then run the Proposition 8.3 construction for the triangle whose outer edge is x=10, using ρ = 10/∥z_{S'}∥, and test whether the image of Ξ_1(T) lies in U = st(V)\\T. If the image leaves U, the central recursive estimate of Theorem 9.3 is not supported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing geometric engine is Proposition 8.3, whose proof invokes Proposition 8.2: st_T(S) is star-shaped with respect to B_{h/(ℓ+1)}(z_S), with h defined as the *maximum* height of any vertex of S in any n-simplex of the star. But the proof of Proposition 8.2 only establishes the needed containment for ϱ at most the *minimum* height of z_S in any n-simplex of st_T(S). For ℓ=0 these differ. Example: take a 2D vertex star formed by fan-triangulating a square with a deep triangular notch; V=(0,0) lies in the kernel, the distance to the closest boundary/notch edge is about 2.1, while the distance to the outer square edge is 10. Then h=10 and B_10(V) contains points outside the notch, so Proposition 8.2 is false as stated. Proposition 8.3 then chooses y with ρ ≤ h/∥z_{S'}∥ and needs y∈B_ϱ(z_S)⊆st_T(S) together with the convex hull of T and y contained in the star. With h taken as maximum this premise can fail, so the recursive estimate in Theorem 9.3 is not derived from valid assumptions. The defect is a min/max inconsistency, not an obscure geometric subtlety, and it directly controls the computable constants that constitute the paper's central contribution. If 'maximum' is replaced by 'minimum', the argument may be repairable, but the constants and all subsequent bounds in Section 9 would need to be recomputed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs linear potential operators for the exterior derivative on two classes of domains: bounded convex domains (regularized Poincaré and Bogovski operators, Theorem 6.2) and domains admitting a shellable simplicial triangulation (Theorems 9.3 and 9.4). The operator norms are bounded by explicitly computable constants depending only on the geometry, yielding upper bounds for Poincaré–Friedrichs constants and lower bounds for vector-Laplacian eigenvalues. Numerical experiments in Section 10 compare the bounds with finite-element reference values on several 2D and 3D examples.","tokens_in":58291,"tokens_out":9567,"duration_ms":115333,"significance":"If the geometric estimates in Section 8 are valid, the shellable-triangulation result is a significant advance: it extends computable Poincaré–Friedrichs constants to curl and divergence operators and to the full Lp scale, including non-convex local patches around reentrant corners, slits, and crossed bricks. The convex-domain part, Theorem 6.2, is a self-contained and parameter-free derivation with explicit constants; it appears sound and has independent value. The numerical examples give an honest assessment of the overestimation factors. However, the central shellable-triangulation claim is not established as written: Proposition 8.2 is false in the stated form, and the main theorem depends on it. The defect is load-bearing but appears repairable within the paper's framework.","major_comments":[{"comment":"The statement of Proposition 8.2 defines h as the maximum height of any vertex of S in any n-simplex of st_T(S), but the proof only supports the radius ϱ at most the minimum height of z_S. In the ℓ=0 case the proof explicitly assumes 'ϱ is at most the minimum height of z_S in any n-simplex'; the containment B_ϱ(z_S)⊆|st_T(S)| can fail for the maximum height. For example, a fan triangulation of a square with a deep notch around an interior vertex can have distance to the closest boundary/notch edge much smaller than the distance to the outer square edge; with h taken as the maximum, B_h(z_S) leaves the star. Proposition 8.3 then requires y∈B_{h/(ℓ+1)}(z_S)⊆st_T(S) and uses the convex hull of T and y to define the reflection; with h as maximum, this premise fails. Consequently Theorem 9.3 is not derived from valid assumptions. Replacing 'maximum' by 'minimum' may repair the argument, but t","section":"Section 8, Lemma 8.1"},{"comment":"Lemma 8.1 is stated without proof, yet it controls the volume relation vol(T')=vol(T)/(ℓ+1), the height-vector relations used in the reduction of Proposition 8.2 to ℓ=0, and the height/Jacobian bounds in Proposition 8.3. Since these relations are directly used in the proof of the reflection estimates, the omission is load-bearing. Please supply a complete proof or a precise reference.","section":"Section 8, Lemma 8.1"}],"minor_comments":[{"comment":"The line 'w = P_kdu' is inconsistent with the indexing P_k:LpΛ^k(Ω)→WpΛ^{k-1}(Ω). For u∈WpΛ^k(Ω), the correct potential is w=P_{k+1}du, as follows from (56). Please correct the indexing.","section":"Theorem 6.2, after equation (56)"},{"comment":"The symbol w''_m is used in 'wm := ewm + w''_m' but is never defined. The subsequent norm estimate indicates the intended construction is wm = Ξ_1^*(w_{m-1}|Um−1) + u|Tm − Ξ_1^*(u|Um−1), i.e., w''_m = u|Tm − Ξ_1^*(u|Um−1). Please clarify or correct the recursion formula.","section":"Theorem 9.3, proof"},{"comment":"The phrase 'maximum height of any vertex of S within any n-simplex of st_T(S)' is ambiguous when dim S>0, because the proof concerns heights of the barycenter z_S. The quantity h should be defined explicitly in terms of the heights of z_S (or of the relevant vertices) in the n-simplices of the star.","section":"Section 8, Proposition 8.2"}],"recommendation":"major_revision","confidential_remarks":"The reader's stress-test concern is correct and lands on the main theorem: Proposition 8.2 is false as stated, and Proposition 8.3/Theorem 9.3 inherit the failure. The convex-domain part (Theorem 6.2) appears sound and publishable independently. I recommend major revision rather than reject because the min/max defect is local in structure and may be repairable, provided the authors replace the maximum by a minimum and recompute the affected constants."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThe headline: this paper is worth a serious referee, but the main theorem for shellable triangulations has a load-bearing flaw. The convex-domain part stands on its own and is genuinely new.\n\nWhat the paper does well: Theorem 6.2 gives explicit, p-independent operator-norm bounds for regularized Poincar\\'e and Bogovski potentials over bounded convex domains, for all form degrees. That is clean and computable, and it fills a real gap. The sequential construction for gradients is known, but carrying it to the full exterior derivative via shellability is a good idea, and the class of shellable complexes correctly includes local stars in 2D/3D.\n\nThe soft spot is Section 8. Proposition 8.2 claims the star st_T(S) is star-shaped with respect to a ball of radius h/(\\ell+1), where h is the maximum height. The proof only supports the minimum height; with the maximum, the ball can stick outside the star (the notch example is convincing). Proposition 8.3 then uses that ball to choose the reflection point y and to ensure the convex hull stays inside the star, so the recursive estimate in Theorem 9.3 is not justified from valid assumptions. Lemma 8.1 is also stated without proof, and it is central to the Jacobian estimates. If the min/max is corrected and Lemma 8.1 supplied, the argument may be repairable, but the constants in Section 9 would need recomputation.\n\nI also want to be fair: the numerical experiments are honest, even though the 3D constants are extremely loose (ratios in the thousands), and the authors say so. No code was deposited, but the algorithms are described well enough to reproduce.\n\nWho benefits from this paper? Numerical analysts working on guaranteed eigenvalue bounds or a posteriori estimates for curl/divergence problems. They should not use Theorem 9.3 as it stands, but the convex results in Section 6 are ready to use.\n\nMy recommendation: send it to peer review, with the expectation that the authors fix Proposition 8.2/8.3 or state the theorem under a corrected geometric assumption. This is real work, but it is not ready as is.","headline":"A real advance in computable constants for the Lp de Rham complex, but the shellable-triangulation part rests on a geometric lemma with a min/max error that breaks the proof of Theorem 9.3 as written.","tokens_in":58666,"tokens_out":2076,"would_cite":false,"duration_ms":24875,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65N30","35P15","58A10"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper constructs bounded potentials for the gradient, curl, and divergence over domains with shellable triangulations, with explicitly computable constants that yield upper bounds on Poincaré–Friedrichs constants and lower bounds on ve","keywords":["Poincaré–Friedrichs constants","Lp de Rham complex","shellable triangulations","potential operators","exterior derivative","computable bounds","Maxwell eigenvalues","convex domains"],"falsifier":"Verify Lemma 8.1 explicitly on a tetrahedron split by barycentric subdivision of an edge: check that each sub-simplex has volume $\\mathrm{vol}(T)/(\\ell+1)$ and that the claimed height-vector scalings hold. Then, on a non-convex boundary star such as the slit-domain patch, compute the actual Jacobian singular values of the piecewise affine reflection $\\Xi_1$ and compare them with the stated $C_{5,n,\\ell}$ and $C_{6,n,\\ell}$ bounds; a violation of either would break the recursive estimate in Theorem 9.3.","tokens_in":57741,"feed_emoji":"📐","tokens_out":6713,"duration_ms":74302,"temperature":0.7,"pith_summary":"The paper establishes that on a domain with a shellable triangulation, every differential form $u \\in W^p\\Lambda^k(\\Omega)$ whose exterior derivative $du$ is in $L^p$ admits a potential $w \\in W^p\\Lambda^k(\\Omega)$ with $dw = du$ and $\\lVert w\\rVert_{L^p(\\Omega)} \\le C(\\mathcal{T},p,k)\\lVert du\\rVert_{L^p(\\Omega)}$, where the constant is explicitly computable from mesh geometry. For bounded convex domains, the paper gives analogous explicit constants for the entire $L^p$ de Rham complex, independent of the Lebesgue exponent $p$ but depending on dimension and form degree. These operator norms are upper bounds for Poincaré–Friedrichs constants and, in the Hilbert case $p=2$, translate directly into lower bounds on eigenvalues of scalar and vector Laplacians, including Maxwell eigenvalues. The gradient, curl, and divergence operators are all treated as instances of the exterior derivative, so the argument covers them uniformly in two and three dimensions.","feed_headline":"Explicit stability constants for curl and divergence on meshes","feed_subtitle":"Shellable triangulations yield provable Poincaré–Friedrichs bounds and lower bounds on vector Laplacian eigenvalues.","key_machinery":"The argument runs on three objects. First, a shellable triangulation: an ordering of the top-dimensional simplices such that each new simplex meets the union of the previous ones in a non-empty union of codimension-one faces, so that gluing can proceed along whole faces rather than along edges or vertices. Second, regularized Poincaré and Bogovskii-type integral operators: averaged potential operators whose Lebesgue-space operator norms give the convex-domain constants of Theorem 6.2. Third, the bi-Lipschitz reflection $\\Xi_1$ of Proposition 8.3, a piecewise affine map from a simplex into the complement of that simplex within its local star, identity on the interface, with explicit bounds on","core_discovery":"The central claim is that constructive potentials for the exterior derivative exist over domains with shellable triangulations, with operator norms bounded by explicit functions of the triangulation's shape measures, volume ratios, and local star geometry. The construction is inductive: starting from one simplex, each new simplex in a shelling is attached along a union of faces, and a potential on the new simplex is built by pulling back the already-constructed potential from the complement of the simplex inside the completed local star, using a bi-Lipschitz piecewise affine reflection that is the identity on the interface. The recursive norm estimates are then unfolded into a global Poincar","pith_inferences":["Beyond the paper: the same gluing philosophy should extend to shellable polytopal complexes, and the paper explicitly suggests lumping simplices into polytopal subdomains to counteract the growth of constants with mesh size.","Beyond the paper: the numerical examples show overestimation factors that explode with the number of simplices, especially in 3D; this suggests that choosing a shelling that optimizes geometric constants is not merely algorithmic convenience but essential for practical usefulness.","Beyond the paper: if the conjectured domination of all $L^p$ de Rham constants by the gradient constant holds for convex domains, then the convex-domain bounds of Theorem 6.2 could be sharpened substantially, and the same sharpening would propagate into the shellable-triangulation estimates."],"forward_implications":["For any shellable triangulation, gradient, curl, and divergence potentials have computable stability constants obtained from local mesh data, with no need to solve global finite element eigenvalue problems.","For $p=2$, the computed upper bounds on Poincaré–Friedrichs constants become lower bounds on Neumann and Dirichlet Laplacian eigenvalues and on Maxwell eigenvalues over the same domain.","Local patches (stars) in two- and three-dimensional triangulations are shellable, so finite element vertex, edge, and face patches now admit curl and divergence stability constants where only gradient constants were previously available.","For bounded convex domains, the constants cover the full $L^p$ de Rham complex and all $p \\in [1,\\infty]$, including partial boundary conditions on simplices, which feeds the recursive construction and is of independent interest.","The potential operators preserve polynomial differential forms, so the construction is compatible with finite element exterior calculus spaces and can be embedded directly into computational pipelines."],"supporting_citations":[{"why":"Supplies the regularized Poincaré and Bogovskii integral operator framework that Section 6 simplifies and whose operator norms are estimated.","marker":"[22]"},{"why":"Provides the definition and background theory of shellable simplicial complexes that the main triangulation results build on.","marker":"[71]"},{"why":"Supplies the Sobolev space, Piola transform, and Poincaré constant background used throughout the paper's notation and estimates.","marker":"[24]"},{"why":"Contributes the local extension and gluing technique in vertex patches that the shellable-star construction adapts to differential forms.","marker":"[25]"},{"why":"Provides the constrained and unconstrained local reconstruction framework in the de Rham complex that motivates the reflection-based potential construction.","marker":"[19]"},{"why":"Gives the earlier computable Poincaré constants for finite element stars in the scalar case, which the present paper extends to curl and divergence.","marker":"[66]"},{"why":"Supplies pullback and transformation estimates and the crossed-bricks domain example used in the numerical study.","marker":"[50]"},{"why":"Provides the $L^p$ change-of-variables inequalities for differential forms used in Proposition 5.3 to bound pullback norms.","marker":"[64]"}],"fun_headline_variants":["Shellable meshes yield explicit Poincaré–Friedrichs constants","Shellable triangulations give computable Poincaré–Friedrichs bounds","Computable constants for vector Laplacians from shellable meshes","Constructive potentials yield explicit bounds for curl, divergence, and eigenvalues","Explicit stability constants on shellable triangulations for vector calculus"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The load-bearing premise is the geometric reflection estimate in Proposition 8.3, whose proof rests on Lemma 8.1, stated without proof, and on the claim that every local star is star-shaped with respect to a ball of radius $h/(\\ell+1)$; if those geometric facts fail, the recursive bound collapses.","fun_headline_variants_meta":{"raw":{"variants":["Shellable meshes yield explicit Poincaré–Friedrichs constants","Shellable triangulations give computable Poincaré–Friedrichs bounds","Computable constants for vector Laplacians from shellable meshes","Constructive potentials yield explicit bounds for curl, divergence, and eigenvalues","Explicit stability constants on shellable triangulations for vector calculus"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001067,"raw_usage":{"total_tokens":4291,"prompt_tokens":711,"completion_tokens":3580,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":3487}},"tokens_in":455,"tokens_out":3580,"duration_ms":29206,"temperature":1.0,"reasoning_tokens":3487,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T22:35:14.028175+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Verify Lemma 8.1 explicitly on a tetrahedron split by barycentric subdivision of an edge: check that each sub-simplex has volume $\\mathrm{vol}(T)/(\\ell+1)$ and that the claimed height-vector scalings hold. Then, on a non-convex boundary star such as the slit-domain patch, compute the actual Jacobian singular values of the piecewise affine reflection $\\Xi_1$ and compare them with the stated $C_{5,n,\\ell}$ and $C_{6,n,\\ell}$ bounds; a violation of either would break the recursive estimate in Theorem 9.3.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the definition and background theory of shellable simplicial complexes that the main triangulation results build on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the constrained and unconstrained local reconstruction framework in the de Rham complex that motivates the reflection-based potential construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the earlier computable Poincaré constants for finite element stars in the scalar case, which the present paper extends to curl and divergence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies pullback and transformation estimates and the crossed-bricks domain example used in the numerical study."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the $L^p$ change-of-variables inequalities for differential forms used in Proposition 5.3 to bound pullback norms."}],"review_version":1}