{"id":"304bdc47-b945-4a82-aa3b-1fab2adfa45c","arxiv_id":"2506.14639","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For real elliptic equations on domains with Ahlfors regular boundaries, the Poisson-Dirichlet problem with Besov boundary data is well posed in a wide range of fractional smoothness spaces.","lead":"This math paper announces new well-posedness results for the Poisson-Dirichlet problem with fractional-smoothness boundary data on very rough domains. It works for arbitrary real elliptic coefficients and Ahlfors regular boundaries, a much broader class than the Lipschitz domains treated before.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Boundary De Giorgi-Nash estimate (21), cited to [DFM21] and not proved here, controls the width of the pentagon via a≥max(α,1-d); if it fails or α=0, Theorem 36 collapses.","rationale":"The reader's weakest-assumption analysis identifies the right soft spot. The paper is an announcement, so the absence of proofs is expected, but the central theorem's range is quantitatively controlled by a cited, unexamined estimate. If (21) holds with α>0, the qualitative claim is plausible; if not, the whole pentagon construction fails. I do not find an internal inconsistency in the stated geometry: the convexity argument in Remark 42 checks out using a*≥1-d, and the Besov definitions are self-consistent, though the paper itself flags that the atomic characterization for p<1 is not found in the literature and is deferred to [BMPa]. That is a further reason for conditional acceptance, but it is secondary to (21), which is load-bearing. The recommendation CONDITIONAL remains appropriate: accept once the companion papers provide complete proofs and the citation to [DFM21] is confirmed to cover the exact class of operators and domains used here.","tokens_in":23867,"tokens_out":19441,"duration_ms":220830,"concrete_test":"Check the statements of [DFM21, Lemmas 8.13 and 8.16] against the hypotheses of Theorem 36: verify that they apply to all real, not necessarily symmetric, matrices A satisfying (15)-(16) on connected d-Ahlfors regular domains with 0<d<n-1, and that they yield (21) with a positive exponent α. If the lemmas are restricted to a subclass of operators or domains, or if α is not positive in the stated generality, then Theorem 36 must be revised or supplied with a proof of (21) under its exact hypotheses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative well-posedness region in Theorem 36 is governed by the boundary De Giorgi-Nash estimate (21) through the relations a≥max(α,1-d) and a*≥max(α*,1-d), and Theorem 45 inherits this dependence through (46). The manuscript does not prove (21) for the full class of real elliptic A satisfying (15)-(16) on d-Ahlfors regular boundaries with 0<d<n-1; it cites [DFM21, Lemmas 8.13 and 8.16]. If those lemmas impose extra hypotheses (e.g., symmetry of A, stronger geometric assumptions, or a different notion of vanishing boundary values), or if the Hölder exponent α can be zero for some admissible operator, then the pentagon in Figure 3 and the hexagons in Figure 4 would shrink or become empty. Because the proof of Theorem 36 is deferred entirely to the companion papers [BMPa] and [BMPb], this manuscript contains no way to verify the mechanism that turns boundary Hölder regularity into the full (s,1/p) range. This is the most load-bearing point: every other ingredient (Meyers estimates, extension results, interpolation) is either standard or auxiliary, whereas this single estimate controls the size of the claimed region.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper announces well-posedness results for the Poisson-Dirichlet problem with boundary data in homogeneous Besov spaces on domains whose boundary is d-Ahlfors regular, for 0<d≤n−1. Theorem 36 asserts that for every real elliptic operator satisfying the degenerate ellipticity conditions (15)–(16), there is a pentagonal range of smoothness/integrability parameters (s,1/p) for which the problem has a unique solution satisfying a weighted averaged gradient estimate; the quality of the range is tied to the exponent α in the boundary De Giorgi–Nash estimate (21). Theorem 45 expands this range to hexagons whenever an endpoint L^q-Dirichlet problem for L or L* is well posed. The manuscript also provides a historical survey of prior work and several corollaries obtained by combining Theorem 45 with known L^q solvability results. The full proofs are not included and are explicitly deferred to two companion papers in preparation.","tokens_in":24162,"tokens_out":9431,"duration_ms":92333,"significance":"If the announced theorems are correct, they constitute a substantial advance: they appear to be the first results for the Poisson-Dirichlet problem with fractional-smoothness Besov boundary data in the full generality of Ahlfors regular boundaries (including lower-dimensional boundaries) and arbitrary real elliptic coefficients with no additional coefficient regularity. The extrapolation mechanism of Theorem 45, which uses only an endpoint L^q-Dirichlet solvability assumption, is elegant and promises broad applicability to the many known L^q solvability results surveyed in Section 7. The paper is well organized, with carefully stated definitions and a useful historical account. Its main weakness is evidentiary: the central theorems are announced without proof, and the key quantitative ingredients, especially the boundary De Giorgi–Nash estimate (21), are cited from other works rather than established or even discussed at the level of hypotheses. The reader cannot currently verify the main claims.","major_comments":[{"comment":"The main theorems are stated without proof. The text explicitly says the full proofs will appear in the in-preparation manuscripts [BMPa] and [BMPb], and the only derivation-like content in the paper is standard material from the literature. Since Theorem 36 and Theorem 45 are the central claims of the paper, and since the estimates (38), (39), (43), and the parameter relations (46) are all products of the deferred arguments, a referee cannot verify that the statements are correct, that the constants a, a*, b, b*, ε, δ are actually available, or that uniqueness holds in the asserted classes. This is a load-bearing incompleteness rather than a presentation issue.","section":"Sections 4 and 5, Theorems 36 and 45"},{"comment":"The size of the well-posedness region in Theorem 36 is controlled by the boundary De Giorgi–Nash estimate (21) through the relations a≥max(α,1−d) and a*≥max(α*,1−d), and Theorem 45 inherits this via (46). The manuscript does not prove (21) for the full class of real elliptic matrices satisfying (15)–(16) on d-Ahlfors regular boundaries; it cites [DFM21, Lemmas 8.13 and 8.16]. If those lemmas require additional hypotheses on A or Ω, or if α can be zero for some admissible operator, the pentagon in Figure 3 and the hexagons in Figure 4 would shrink or become empty. Because the proof is deferred, the manuscript provides no way for the reader to verify the mechanism by which boundary Hölder regularity is converted into the full claimed (s,1/p) range.","section":"Section 2.4, estimate (21), and the relations a≥max(α,1−d), a*≥max(α*,1−d)"},{"comment":"The definition of the atomic Besov spaces ˙A^{s,p}(Γ) relies on the assertion that finite block sums converge in ˙Λ^{s+d−d/p,1}(Γ) and that ˙A^{s,p}(Γ) embeds there. The text states this will be proved in [BMPa] and notes that the atomic characterization appears not to be in the literature. Since these spaces are used to define the boundary data spaces ˙B^{p,p}_s(∂Ω) for p<1, the absence of a proof is another load-bearing gap: the spaces in which boundary values are taken are not fully established within this manuscript.","section":"Definition 10 and Section 2.2"}],"minor_comments":[{"comment":"The algebraic condition (41) appears to contain a typo: for the pentagon in Figure 3, which has vertex at the origin and edge from (0,0) to (1−a*,1), the lower bound should be (1−a*)/p < s, not 1−a*/p < s. As printed, (41) gives no admissible s for p=∞, contradicting the theorem's own statement that 0<s<a is allowed for p=∞.","section":"Theorem 36, paragraph after (41)"},{"comment":"The vertices of the pentagon would be clearer if the figure explicitly marked the point (0,0) and the edge corresponding to p=∞, 0<s<a; the current caption and the algebraic conditions do not make the relationship immediate.","section":"Figure 3 and Remark 42"},{"comment":"Several references are to works without complete publication data, including [CHPM+], [Bar], [Fen], and [MPT]. Since the proof relies on [DFM21], it would be helpful to indicate the exact statements and hypotheses of Lemmas 8.13 and 8.16, preferably by quoting them.","section":"References"},{"comment":"The statement that Ahlfors regularity for d=n−1 implies the Wiener-type criterion of [CHPM+] is plausible but is asserted without a reference or argument; a citation or a one-sentence explanation would improve the exposition.","section":"Section 3.9"}],"recommendation":"major_revision","confidential_remarks":"The main issue is not a suspected error but the absence of proofs for the central theorems. If the journal's policy permits research announcements, the editor may decide that deferred proofs are acceptable provided the companion papers are close to complete; otherwise, the manuscript needs to include the proofs or a much more detailed account of the key estimates. I recommend requiring the authors to either supply the full proofs, including the verification of the boundary De Giorgi–Nash estimate under exactly the hypotheses of Theorem 36, or to make the companion manuscripts publicly available and cite them with precise references to theorem numbers."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: read this if you want the current frontier on Poisson-Dirichlet with rough coefficients. It announces the first Besov-space well-posedness results for arbitrary real elliptic A in Ahlfors regular domains, including higher codimension, without DKP or t-independence assumptions. The historical survey is genuinely useful and the corollaries tying the main theorem to known Lq solvability are stated cleanly.\n\nWhat's good: the two main theorems are precise. Theorem 36 gives a pentagonal (s,1/p) region controlled by De Giorgi-Nash and Meyers exponents; Theorem 45 expands it using endpoint Lq solvability. The paper is honest about what is new and what is not, and it flags its own limitations: n=2 open, d>n-1 open, proofs in companions.\n\nSoft spots: every central theorem is stated without proof. The only derivation-like content is standard material. That is acceptable for an announcement, but it means a referee cannot verify the key mechanism. The most load-bearing external piece is the boundary De Giorgi-Nash estimate (21), cited to [DFM21, Lemmas 8.13 and 8.16]. The pentagon in Figure 3 and hexagons in Figure 4 shrink or collapse if α=0 or if those lemmas impose extra hypotheses beyond (15)-(16). My reading of [DFM21] is that the citation is plausible—it is a published Memoir and the conditions look aligned—but this paper does not do the verification, so the reader is stuck. Also deferred is the atomic characterization of homogeneous Besov spaces on spaces of homogeneous type; that is a smaller issue, but it is a missing proof for a definition they rely on.\n\nFor whom: people working in elliptic boundary value problems with rough data, or anyone needing a map of the fractional-smoothness landscape. It deserves a serious referee, but the referee should have access to [BMPa] and [BMPb], or at least should be willing to check the cited [DFM21] lemmas against the ellipticity hypotheses. My recommendation: engage with it, send it to a referee, but tie acceptance to the companion proofs or evidence that (21) holds for the full class of operators considered. I would not cite the main results in my own work until those proofs appear, but I would keep this as the standard reference for the current landscape.","headline":"A careful, genuinely new announcement of Besov-space well-posedness for rough elliptic operators on Ahlfors regular boundaries, whose central claims rest on companion papers and one load-bearing cited estimate that the reader cannot verify here.","tokens_in":24629,"tokens_out":2930,"would_cite":false,"duration_ms":29533,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35J25","35A01","35A02"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes well-posedness of the Poisson–Dirichlet problem with Besov boundary data of fractional smoothness on domains with Ahlfors regular boundary, for arbitrary real elliptic coefficients and no layer-potential machinery.","keywords":["Poisson-Dirichlet problem","Besov spaces","fractional smoothness","Ahlfors regular boundary","higher codimension","real elliptic operators","weighted Sobolev estimates","well-posedness"],"falsifier":"Construct, or find in the literature, an admissible domain and real elliptic matrix satisfying (15)–(16) for which the exponent in the boundary De Giorgi–Nash estimate (21) is zero; then the relations $a\\ge\\max(\\alpha,1-d)$ and $a^*\\ge\\max(\\alpha^*,1-d)$ force the pentagon (40)–(41) to have empty interior, contradicting Theorem 36. Alternatively, for the Laplacian in a weak local John domain with $(n-1)$-Ahlfors regular boundary, take $(s,1/p)$ just outside the claimed region and seek a sequence of Besov data with bounded norm whose solutions violate (38); a proven violation fixes the sharp boundary, while uniform boundedness would suggest the true region is larger.","tokens_in":23683,"feed_emoji":"📐","tokens_out":20642,"duration_ms":171503,"temperature":0.7,"pith_summary":"This paper announces a proof that the Poisson–Dirichlet problem is well posed with boundary data in Besov spaces of fractional smoothness on domains far rougher than Lipschitz: connected domains whose boundary is $d$-Ahlfors regular for $0<d\\le n-1$, with the weak local John and interior corkscrew conditions added when $d=n-1$, and whose coefficients are merely real, bounded, measurable, and elliptic in the natural distance-weighted sense. Previously known results in this range of smoothness were mostly confined to Lipschitz domains with VMO, Dini, or transversally constant coefficients. The proof is 'blind': it uses only the boundary De Giorgi–Nash estimate, Meyers's reverse Hölder estimate, and interpolation and duality, so it requires no layer potentials, no condition controlling coefficient oscillation near the boundary (DKP), and no assumption that coefficients be constant in the transverse direction. Moreover, if an endpoint $L^{q}$-Dirichlet problem is known for the operator or its adjoint, the range of admissible pairs $(s,1/p)$ automatically expands to a larger hexagon. If correct, this makes fractional-smoothness Dirichlet data a routine consequence of real ellipticity plus mild boundary geometry.","feed_headline":"Besov data solve Poisson-Dirichlet on Ahlfors-regular boundaries","feed_subtitle":"Needs no smooth coefficients or Lipschitz domains — only an Ahlfors-regular boundary and real ellipticity.","key_machinery":"The load-bearing estimate is the boundary De Giorgi–Nash inequality (21): if $-\\operatorname{div} A\\nabla u=0$ in $\\Omega$, $u=0$ on $\\partial\\Omega\\cap B(\\xi,r)$, and $x\\in B(\\xi,r/2)\\cap\\Omega$, then $|u(x)|\\le C(|x-\\xi|/r)^\\alpha r^{-(d+1)}\\int_{\\Omega\\cap B(\\xi,r)} |u(y)|\\operatorname{dist}(y,\\partial\\Omega)^{1+d-n}\\,dy$. It supplies the positive Hölder exponent $\\alpha$ that converts energy-level solvability into a range of fractional Besov exponents, and the widths $a,a^*,b,b^*$ are controlled by $\\alpha,\\alpha^*$ through (40)–(41) and (46). The second mechanism is Meyers's reverse Hölder estimate (20), which gives the $\\beta>2$ integrability of gradients used in the averaged norm in (38). For Theorem 45, the endpoint ingredient is the nontangential estimate (43) for the $L^{q}$-Dirichlet problem, so that known endpoint results transfer automatically to the larger hexagon without any further conditions on coefficients or geometry.","core_discovery":"The central claim is Theorem 36: for $n\\ge 3$ and a connected open set $\\Omega$ whose boundary is $d$-Ahlfors regular for some $0<d\\le n-1$ (with $\\Omega$ weak local John and interior corkscrew when $d=n-1$), and for every real elliptic matrix $A$ satisfying (15)–(16), the Poisson–Dirichlet problem\n$$-\\operatorname{div} A\\nabla u=-\\operatorname{div}(A\\vec H)\\quad\\text{in }\\$\\Omega$,\\qquad u=f\\quad\\text{on }\\partial\\$\\Omega$$$\nhas a unique solution for every $f\\in\\dot B^{p,p}_s(\\partial\\Omega)$ and every admissible $\\vec H$, whenever $(s,1/p)$ lies in the open pentagon described by (40)–(41), with the weighted averaged estimate (38); the $p=\\infty$, $0<s<a$ endpoint gives (39). The numbers $a,a^*$ obey $a\\ge\\max(\\alpha,1-d)$ and $a^*\\ge\\max(\\alpha^*,1-d)$, where $\\alpha,\\alpha^*$ are the boundary De Giorgi–Nash exponents of $L$ and $L^*$ from (21), and the integrability exponent $\\beta\\in(2-\\delta,2+\\varepsilon)$ is supplied by Meyers's reverse Hölder estimate (20). Theorem 45 then shows that if the $L^{q}$-Dirichlet problem for $L$ (or the $L^{q^*}$-Dirichlet problem for $L^*$) is well posed in the sense of the nontangential estimate (43), the pentagon expands to the larger hexagons of Figure 4, with $b\\ge\\alpha$, $b^*\\ge\\alpha^*$, $b\\ge 1+d/q^*-d$, and $b^*\\ge 1+d/q-d$. The announcement emphasizes that no condition beyond real ellipticity is imposed on the coefficients, and that the higher-codimension case $d<n-1$ automatically satisfies the extra geometric hypotheses.","pith_inferences":["A natural extension is the $n=2$ case: the paper's stated obstruction is tied to fundamental-solution behaviour in two dimensions, so the theorem might hold there via a different estimate, or the dimension may be a genuine limitation.","If the boundary De Giorgi–Nash estimate (21) were available with a positive exponent for complex coefficients or elliptic systems, the same interpolation mechanism would likely produce an analogous pentagon; the present theorem is restricted to real coefficients.","The extrapolation in Theorem 45 points to a general transfer principle: every new endpoint $L^{q}$-Dirichlet solvability result for a rough coefficient class automatically buys a region of fractional Besov well-posedness, shifting the bottleneck of the theory to the endpoint problem.","Because $a^*\\ge 1-d$ and $b^*\\ge 1+d/q-d$, the usable region narrows as the boundary dimension $d$ decreases; for very low-dimensional boundaries the geometric term $1-d$ can dominate the analytic exponent $\\alpha$ — a quantitative prediction that could be checked in examples."],"forward_implications":["Besov boundary data of fractional smoothness are handled on $d$-Ahlfors regular domains, including higher-codimension boundaries, for all real elliptic operators satisfying (15)–(16); no VMO, Dini, DKP-type oscillation condition, or transverse-independence assumption is needed.","In the codimension-one case, once the $L^{q}$-Dirichlet problem is known for $L$ or $L^*$, Theorem 45 automatically enlarges the Besov range to the hexagons in Figure 4, so endpoint progress directly improves the fractional-smoothness theory.","For the Laplacian on weak local John domains with uniformly rectifiable $(n-1)$-Ahlfors regular boundary, the known endpoint $L^{q}$-solvability yields the full hexagon at the bottom of Figure 4.","The $p=\\infty$, $0<s<a$ case supplies Hölder-continuous boundary data and the sup-norm estimate (39), matching the known Hölder well-posedness on the $s\\to 0$ edge.","The Besov norm on the boundary data is the sharp trace norm for the weighted averaged gradient space in (38), so the right-hand side in $f$ cannot be weakened."],"supporting_citations":[{"why":"Supplies the boundary De Giorgi–Nash estimate (Lemmas 8.13 and 8.16) and the higher-codimension extension/trace framework that the pentagon theorem builds on.","marker":"[DFM21]"},{"why":"Supplies the boundary trace definition (13) and the Besov-to-weighted-Sobolev extension theorem that makes u=f on the boundary meaningful.","marker":"[JW84]"},{"why":"Cited as the forthcoming full proof of the Dirichlet problem with fractional-smoothness boundary data, including the extension results used in Section 2.","marker":"[BMPa]"},{"why":"Cited as the forthcoming full proof of the Poisson problem and the interpolation argument behind Theorem 36.","marker":"[BMPb]"},{"why":"The prior weighted-Sobolev Besov result in Lipschitz domains with Dini-type coefficients that Theorem 36 generalizes to rough domains.","marker":"[MT06]"},{"why":"The prior VMO-coefficient result in Lipschitz domains, another baseline that the new 'blind' method replaces.","marker":"[MMS10]"},{"why":"Supplies the endpoint L^q-Dirichlet solvability for the Laplacian on weak local John domains with uniformly rectifiable boundary, activating Theorem 45 in codimension one.","marker":"[AHM+20]"},{"why":"Supplies the endpoint absolute-continuity result for low-dimensional uniformly rectifiable boundaries, activating Theorem 45 in higher codimension.","marker":"[DM22]"},{"why":"States the equivalence between L^q-Dirichlet solvability and reverse Hölder estimates for harmonic measure, used to turn the geometric and coefficient hypotheses of Section 7 into endpoint well-posedness.","marker":"[HL18]"}],"fun_headline_variants":["Poisson-Dirichlet works on Ahlfors-regular boundaries with Besov data","Besov Poisson-Dirichlet on rough boundaries: no smooth coefficients","Ahlfors-regular domains solve Poisson-Dirichlet for fractional Besov data","Poisson-Dirichlet for Besov data on Ahlfors-regular boundaries, no Lipschitz","Real elliptic coefficients suffice for Poisson-Dirichlet on Ahlfors-regular sets"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire fractional-smoothness range rests on the boundary De Giorgi–Nash estimate (21) holding with a strictly positive exponent for real elliptic operators on $d$-Ahlfors regular domains; if that Hölder decay were absent, or the estimate failed, the pentagons of Figures 3–4 would collapse to zero width.","fun_headline_variants_meta":{"raw":{"variants":["Poisson-Dirichlet works on Ahlfors-regular boundaries with Besov data","Besov Poisson-Dirichlet on rough boundaries: no smooth coefficients","Ahlfors-regular domains solve Poisson-Dirichlet for fractional Besov data","Poisson-Dirichlet for Besov data on Ahlfors-regular boundaries, no Lipschitz","Real elliptic coefficients suffice for Poisson-Dirichlet on Ahlfors-regular sets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000244,"raw_usage":{"total_tokens":1578,"prompt_tokens":1034,"completion_tokens":544,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":650,"completion_tokens_details":{"reasoning_tokens":435}},"tokens_in":650,"tokens_out":544,"duration_ms":4889,"temperature":1.0,"reasoning_tokens":435,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:49:52.441445+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct, or find in the literature, an admissible domain and real elliptic matrix satisfying (15)–(16) for which the exponent in the boundary De Giorgi–Nash estimate (21) is zero; then the relations $a\\ge\\max(\\alpha,1-d)$ and $a^*\\ge\\max(\\alpha^*,1-d)$ force the pentagon (40)–(41) to have empty interior, contradicting Theorem 36. Alternatively, for the Laplacian in a weak local John domain with $(n-1)$-Ahlfors regular boundary, take $(s,1/p)$ just outside the claimed region and seek a sequence of Besov data with bounded norm whose solutions violate (38); a proven violation fixes the sharp boundary, while uniform boundedness would suggest the true region is larger.","supporting_citations":[],"review_version":2}