{"id":"9e0e529d-1b93-4ed2-81c5-6616470500c1","arxiv_id":"2505.23976","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For snakes and circular snakes, greedy pancake decompositions are minimal and canonical up to weakly outer bi-Lipschitz equivalence.","lead":"The paper gives an explicit greedy rule for cutting certain singular surfaces, called snakes and circular snakes, into a minimal number of well-behaved pieces called pancakes. The cut is canonical under weak outer Lipschitz equivalence, which is the right equivalence for the open outer classification problem for surface singularities.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unproved equivalence between non-LNE triangles and non-primitive word segments in Remarks 4.3 and 5.7 is load-bearing: it fixes the greedy cut points that drive Theorems 4.9, 5.3, 5.11 and Proposition 6.1.","rationale":"The reader's verdict is CONDITIONAL and I agree with its weakest_assumption. However, the single most load-bearing issue is narrower than 'classification theorems may have exceptions': the paper changes the definition of the cut points from a geometric predicate to a word predicate in Remarks 4.3 and 5.7, and labels the change as immediate. All of the main theorems either use the greedy cut points to build pancakes (Theorems 4.9, 5.3, 5.11) or compare them under weak equivalence (Propositions 6.1-6.3). If the equivalence is true, the paper's argument is largely sound; if it is false, the minimality claim can fail even when the classification theorems are correct. The dependence on [13] and [10] is real but is an external-input risk rather than an internal gap. I do not see a contradiction or a clear counterexample in the text; the missing lemma is likely provable from Proposition 4.56 and Lemma 4.7 of [13], but the authors should supply it. Hence the verdict should remain CONDITIONAL rather than accept or reject. The concrete enumeration or proof described above would settle the question.","tokens_in":23840,"tokens_out":24395,"duration_ms":234760,"concrete_test":"State and prove the missing lemma: for any a < b in a snake name (resp. circular snake name), T(theta_a, theta_b) is LNE iff [x_a ... x_b] contains no repeated letter. A direct way to test it is to enumerate all snake names of length at most 8 satisfying Definition 6.6 of [13], construct the corresponding snakes via Theorem 6.23 of [13], and for every pair (a,b) compute LNE with the arc criterion of Remark 2.8 (tord = itord for all arcs), comparing the outcome with the repeated-letter test. At minimum, run this comparison on the words in Examples 4.4, 4.5, 5.8, 5.9 and 6.4; if every non-LNE segment contains a repeated letter and every repeated-letter segment is non-LNE, the algorithm's cut points are validated, otherwise the greedy construction must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Remarks 4.3 and 5.7 replace the geometric condition 'T(theta_a, theta_b) is not LNE' in Definitions 4.1 and 5.5 by the combinatorial condition 'the subword [x_a ... x_b] contains a repeated letter'. This is asserted as a direct consequence, but no proof or reference to a supporting lemma is supplied. The greedy cut points are exactly the first positions where this equivalence is invoked, so if the equivalence has any exception -- for instance, a sub-snake whose failure of LNE is caused by cluster multiplicities or node data not visible in the word before the first repeated letter -- then the algorithm computes a different decomposition from the geometric minimal sequence. The minimality lower bound in Theorem 4.9, and its circular analogues, depends on the cut points being precisely the pairs N_{j_{i-1}}, N_{j_i} that cannot lie in a common LNE pancake; an incorrect cut would break that lower-bound argument. Proposition 6.1 inherits the same dependence when it compares greedy decompositions of weakly equivalent snakes. The paper needs a standalone proof of the equivalence, or a citation to an existing lemma in [13] or [10] that establishes it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the outer Lipschitz geometry of 'snakes' and 'circular snakes', abnormal surface germs introduced by Gabrielov and Souza. It defines a 'greedy pancake decomposition' via a minimal sequence of cut points computed from the snake name, and proves three main results: the greedy decomposition is a minimal pancake decomposition for snakes (Theorem 4.9) and for circular snakes with and without nodal zones (Theorems 5.3 and 5.11), and greedy decompositions of weakly outer equivalent snakes or circular snakes are themselves weakly outer bi-Lipschitz equivalent (Propositions 6.1–6.3). The paper also includes examples showing that the greedy algorithm fails for general Hölder triangles (Example 7.2) and that the canonicity does not hold for the stronger notion of outer equivalence (Example 7.3).","tokens_in":24063,"tokens_out":9641,"duration_ms":83797,"significance":"If correct, the paper gives the first constructive, canonical minimal pancake decompositions for snakes and circular snakes, which are key objects in the outer Lipschitz classification problem. The algorithms are explicit and the examples are informative, and the paper properly credits the classification theorems of [13] and [10] on which the canonicity results depend. The main reservation is that the reduction from the geometric non-LNE condition to a combinatorial non-primitivity condition is asserted without proof; since this reduction is the engine of the greedy algorithm, the central claims are not yet fully established.","major_comments":[{"comment":"The equivalence between the geometric condition 'T(θ_a, θ_b) is not LNE' and the combinatorial condition 'the subword [x_a … x_b] contains a repeated letter' is asserted as a 'direct consequence' of Definitions 4.1 and 5.5, but no proof or supporting reference is given. This equivalence is load-bearing: the greedy cut points j_i are exactly the first positions where the equivalence is invoked, and the minimality lower bound in Theorem 4.9 (and its circular analogues) depends on those cut points being precisely the pairs N_{j_{i-1}}, N_{j_i} that cannot lie in a common LNE pancake. If the equivalence has any exception—for example, a sub-snake whose failure of LNE is caused by cluster multiplicities or node data not visible in the word before the first repeated letter—the algorithm may produce different cut points and the lower-bound argument would not apply. The paper needs a standalone proof of this equivalence or an exact citation to a lemma in [13] or [10] that establishes it.","section":"§4 Remark 4.3 and §5 Remark 5.7"},{"comment":"The manuscript asserts that the minimal sequence is independent of the choice of the arcs θ_i in the nodal zones, but no proof is provided. This independence is essential for the well-definedness of the greedy decomposition (Definition 4.10) and for the canonicity statements in Propositions 6.1–6.3. The assertion may be derived from the word-level characterization of Remark 4.3 once that characterization is proved, but as written it is an unproved claim about the geometric definition. Please supply a proof or a precise reference.","section":"§4 Remark 4.2 and §5 Remark 5.6"},{"comment":"Several key steps are deferred with 'follows analogously' or 'admits the same adaptation': the proof that X_i is LNE for i > 2 in Theorem 4.9, the corresponding part of Theorem 5.11, and the treatment of bubbles and spiral snakes in Proposition 6.1. The case tord(S_{j_2}, S_{j_1}) > β in the proof of Theorem 4.9 shows that the choice of λ_i in condition (2) is genuinely delicate, so the remaining cases are not identical to the ones written out. The reader cannot verify without an explicit induction or a clear statement of why the analogous argument applies to every i. Please expand these passages enough to make the inductive step transparent, and spell out the adaptation for bubbles and spiral snakes in Proposition 6.1.","section":"§4 Theorem 4.9; §5 Theorem 5.11; §6 Proposition 6.1"}],"minor_comments":[{"comment":"In the proof of Theorem 5.3 the notation {X_i}_{p+1}^{i=1} is used with p undefined; it should be {X_i}_{m+1}^{i=1} throughout the proof.","section":"§5 Theorem 5.3"},{"comment":"The theorem is stated as '(Theorem 8.3 in [10])', but the statement covers β-snakes as well as circular β-snakes; the snake case is Theorem 6.28 of [13], not Theorem 8.3 of [10]. Please correct the attribution or split the statement according to the source.","section":"§2 Theorem 2.32"},{"comment":"In the proof of Proposition 6.1, the sentence 'θ'_0 = γ'_1, θ'_n = γ_2' should read 'θ'_0 = γ'_1, θ'_n = γ'_2'; the current wording is a typo.","section":"§6 Proposition 6.1"},{"comment":"There are several minor typographical issues, including 'lenght' for 'length' in Remark 2.27, 'bubles' for 'bubbles' in Example 7.1, and 'the the desired weakly outer homeomorphism' in the proof of Proposition 6.1. A careful proofreading pass is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper depends heavily on [10], which is an arXiv preprint by the same group and may not yet be refereed; the editors may wish to verify its status and that of [13]. The critical technical point is the word-primitivity equivalence in Remarks 4.3 and 5.7: if the authors cannot supply a proof or an exact reference, the algorithmic claims and the minimality results are not established. The paper is otherwise well-organized, the examples are helpful, and the reliance on the classification theorems is explicit, so the issues appear fixable within the scope of a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: the paper does something new and useful—gives a greedy algorithm for minimal pancake decompositions of snakes and circular snakes and proves those decompositions are canonical up to weak outer bi-Lipschitz equivalence—but one load-bearing equivalence is asserted without proof, and the referee should ask for it before signing off.\n\nWhat is actually new: minimality of the greedy decompositions (Theorems 4.9, 5.3, 5.11), the fundamental-vs-minimal sequence distinction for circular snakes (Definition 5.5, Example 5.9, Lemma 5.10), and the weak canonicity results (Propositions 6.1–6.3), plus the sharpness examples showing that outer canonicity fails (Example 7.3) and that a naive 'join LNE adjacent pieces' procedure does not give a minimal decomposition in general (Example 7.2). The exposition is careful about what is being assumed from [13] and [10]. I credit the algorithmic clarity: the greedy construction is concrete and the examples (Figures 4, 5, 8, 9) are genuinely helpful.\n\nThe soft spot is real. Remarks 4.3 and 5.7 state, as a 'direct consequence' of the geometric Definition 4.1, that a Holder triangle T(θ_a, θ_b) between nodal zones fails to be LNE exactly when the corresponding snake-name subword [x_a...x_b] is not primitive. No proof or citation is given. This equivalence is not decorative: the greedy cut points j_i are defined as the first place this condition holds. If the word-level condition has any exception—say, a failure of LNE caused by cluster multiplicities that are not visible in the word before the first repeated letter—then the algorithm is computing a different sequence from the geometric minimal sequence, and the lower-bound argument in Theorem 4.9 (and its circular analogues) does not go through. This needs a short proof or an explicit pointer to a lemma in [13] or [10].\n\nThe other soft spots are milder. Theorem 4.9 says the proof for i>2 'follows analogously', and Theorem 5.11 and Proposition 6.1 have similar compressed passages, including the bubble/spiral-snake case at the end of Proposition 6.1. These are probably routine, but in a paper whose main claim is an algorithm, 'follows analogously' is exactly where the reader needs to see the induction or the case split. The reliance on the authors' own classification theorems is not circular, but it concentrates the correctness risk.\n\nBottom line: this is a serious paper for the Lipschitz geometry of surface germs. I would send it to peer review and ask for the missing proof of the word-geometry equivalence and an expansion of the analogous cases. If those come back clean, it is a solid, citable contribution.","headline":"A useful, mostly sound algorithmic paper whose greedy construction leans on an unproved word-geometry equivalence that a referee should ask to be made explicit.","tokens_in":24623,"tokens_out":3271,"would_cite":true,"duration_ms":28502,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14P10","03C64","51F30"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that a greedy word-reading algorithm produces a pancake decomposition with the minimal number of pieces for every snake and circular snake surface germ, and that the decomposition is canonical up to weak outer…","keywords":["pancake decomposition","snake","circular snake","surface germs","outer Lipschitz equivalence","Lipschitz normally embedded","minimal sequence","greedy algorithm"],"falsifier":"Run the greedy algorithm on any concrete snake name, for instance $W=[abacdbcd]$, and check that each pair of adjacent pancakes has a non-LNE union and that the number of pancakes cannot be reduced by merging. More decisively, search the finite set of circular snake names for a fundamental sequence $\\{j_k\\}$ that violates the inequality $p \\geq t(q-1)+1$ from Lemma 5.10; the theory says none exists.","tokens_in":23609,"feed_emoji":"🐍","tokens_out":7579,"duration_ms":70320,"temperature":0.7,"pith_summary":"The paper works on the outer Lipschitz classification of surface singularities, where the basic abnormal pieces are called snakes and circular snakes. It claims that for every such germ there is an algorithmically constructed pancake decomposition with the minimum possible number of pancakes, and that any two weakly outer bi-Lipschitz equivalent snakes or circular snakes have greedy decompositions that are themselves weakly equivalent. The algorithm reads the combinatorial snake name as a word and cuts at the first letter that repeats a previous letter, a step repeated until the word is exhausted. If the theorems are correct, every snake and circular snake carries a canonical minimal decomposition, giving a concrete invariant for the still-open outer classification problem.","feed_headline":"Greedy word cuts give minimal pancake decompositions","feed_subtitle":"For snake-like singularities, the fewest normally embedded pieces come from cutting where a letter repeats.","key_machinery":"The load-bearing object is the snake name, a word whose letters are the nodal zones of the surface, and the greedy rule that reads it. A word segment is primitive exactly when it has no repeated letter; geometrically this corresponds to a normally embedded piece, so the algorithm cuts a surface wherever the accumulated word first repeats a letter. The resulting minimal (or fundamental) sequence supplies the boundary arcs of the pancakes. For circular snakes with nodal zones, the same rule applied to the infinite periodic word can return to its starting zone after more than one lap, and the paper's lifting construction (Lemma 5.10) bounds the period by the number of laps, reducing the circular problem to the snake problem. The choice of the actual arc inside each segment is constrained by a single horn-neighborhood condition that guarantees each H\\\"older triangle is LNE.","core_discovery":"The central discovery is that minimality for snakes and circular snakes is a word-combinatorial phenomenon, not a metric search. For a snake, the greedy cut points are the entries of the minimal sequence $j_0=0$, $j_i=\\min\\{k>j_{i-1} : [x_{j_{i-1}}\\cdots x_k] \\text{ is not primitive}\\}$, and the pieces $T(\\lambda_{i-1},\\lambda_i)$ form a minimal pancake decomposition (Theorem 4.9). The same rule, adapted to the eventually periodic sequence of a circular snake with nodal zones, yields a minimal decomposition (Theorem 5.11), while a circular snake without nodal zones of multiplicity $m$ yields $m+1$ pieces (Theorem 5.3). Propositions 6.1 to 6.3 then show that weak outer bi-Lipschitz equivalence preserves the greedy decomposition, so the minimal number of pancakes and the pattern of cut points are canonical weak invariants. The paper also proves this canonicity is sharp: it fails for the stronger outer equivalence, and the greedy rule need not give minimal decompositions for arbitrary H\\\"older triangles.","pith_inferences":["One testable extension is an exhaustive computer search over snake names and circular snake names of small length, verifying that the greedy cut points always produce reduced decompositions and that the number of pieces equals the minimal count; this would confirm the theorems in a finite range.","The inequality $p \\geq t(q-1)+1$ from Lemma 5.10 may be a purely word-combinatorial statement; if so, it could be detached from surface geometry and checked or generalized in symbolic dynamics.","Since any surface germ is either a circular snake or contains finitely many snakes, the greedy decomposition gives a canonical candidate for the abnormal part of an arbitrary germ; the failure examples for H\\\"older triangles suggest that a complete outer classification will need to record tangency orders within segments, not just the letter multiplicities."],"forward_implications":["Every snake and every circular snake has a computable pancake decomposition of provably minimal size.","The number of pancakes in the greedy decomposition is a weak outer bi-Lipschitz invariant, shared by all weakly equivalent surfaces.","The greedy boundary arcs can be chosen inside segments with a prescribed horn-neighborhood condition, giving explicit cut points rather than an existence argument.","The word-level description makes minimal decompositions checkable by combinatorial computation: any weakly outer bi-Lipschitz map between equivalent snakes must send the greedy cut pattern to the corresponding pattern.","The sharpness examples show that the same canonicity fails under full outer equivalence, so weak equivalence is exactly the level where the decomposition is canonical."],"supporting_citations":[{"why":"Defines snakes and snake names, and provides the weak classification theorem on which the canonicity proofs are built.","marker":"[13]"},{"why":"Introduces circular snakes, their names, and the classification theorems for circular snakes with and without nodal zones.","marker":"[10]"},{"why":"Introduces pancake decompositions and proves that every semialgebraic set has one, which is the existence result the greedy algorithm refines.","marker":"[9]"},{"why":"Establishes the decomposition into normally embedded subsets, the construction later called pancake decomposition, guaranteeing that minimal decompositions exist.","marker":"[15]"},{"why":"Provides the inner bi-Lipschitz classification of surface germs into H\\\"older triangles and horns, which is the geometric language used to define pancakes and snakes.","marker":"[1]"}],"fun_headline_variants":["Greedy cuts give minimal pancakes for snake surface germs","Minimal pancakes via greedy primitive cuts in snakes","Greedy word rule gives canonical minimal pancakes","Snake germs: minimal pancakes from greedy letter cuts","Minimal pancake count is canonically word-combinatorial"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction assumes that a segment of the snake name with no repeated letter exactly matches the geometric fact that the corresponding part of the surface is Lipschitz normally embedded, and that this dictionary between words and geometry has no exceptions beyond the snakes and circular snakes already classified.","fun_headline_variants_meta":{"raw":{"variants":["Greedy cuts give minimal pancakes for snake surface germs","Minimal pancakes via greedy primitive cuts in snakes","Greedy word rule gives canonical minimal pancakes","Snake germs: minimal pancakes from greedy letter cuts","Minimal pancake count is canonically word-combinatorial"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000679,"raw_usage":{"total_tokens":3071,"prompt_tokens":915,"completion_tokens":2156,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":531,"completion_tokens_details":{"reasoning_tokens":2079}},"tokens_in":531,"tokens_out":2156,"duration_ms":14224,"temperature":1.0,"reasoning_tokens":2079,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:39:11.757061+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the greedy algorithm on any concrete snake name, for instance $W=[abacdbcd]$, and check that each pair of adjacent pancakes has a non-LNE union and that the number of pancakes cannot be reduced by merging. More decisively, search the finite set of circular snake names for a fundamental sequence $\\{j_k\\}$ that violates the inequality $p \\geq t(q-1)+1$ from Lemma 5.10; the theory says none exists.","supporting_citations":[{"cited_title":"Gabrielov, E","cited_arxiv_id":null,"evidence_quote":"Defines snakes and snake names, and provides the weak classification theorem on which the canonicity proofs are built."},{"cited_title":"Lipschitz geometry and combinatorics of circular snakes","cited_arxiv_id":"2312.04446","evidence_quote":"Introduces circular snakes, their names, and the classification theorems for circular snakes with and without nodal zones."},{"cited_title":"Birbrair, T","cited_arxiv_id":null,"evidence_quote":"Introduces pancake decompositions and proves that every semialgebraic set has one, which is the existence result the greedy algorithm refines."},{"cited_title":"Kurdyka, P","cited_arxiv_id":null,"evidence_quote":"Establishes the decomposition into normally embedded subsets, the construction later called pancake decomposition, guaranteeing that minimal decompositions exist."},{"cited_title":"Birbrair","cited_arxiv_id":null,"evidence_quote":"Provides the inner bi-Lipschitz classification of surface germs into H\\\"older triangles and horns, which is the geometric language used to define pancakes and snakes."}],"review_version":1}