{"id":"8fe00c64-d8cb-4366-b3e8-8ee068beef44","arxiv_id":"2506.04012","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a gentle algebra, indecomposables in the 1-periodic derived category are exactly string objects from curves between marked points and band objects from primitive closed curves with indecomposable k[x,x^-1]-modules.","lead":"This paper classifies the indivisible building blocks of the 1-periodic derived category of gentle algebras, describing each one by a curve on an associated surface. It gives mathematicians a complete list of indecomposable objects in a category that appears in homological algebra and in Fukaya-category geometry.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 3.3.5(b) is unproved: the map H is never shown to be a morphism of S(Y,σ,k), because discarding blocks outside v=uα can break the commutation relation H(T)G(M)=G(N)H(T); the completeness step of Theorem 3.3.6 rests on this lift.","rationale":"The reader's weakest assumption identifies the lifting step in Proposition 3.3.5(b); I agree that this is the linchpin. My analysis sharpens the gap: the proof's functor H is not even shown to be well-defined on Hom_U, since it discards all blocks whose row is not obtained from the column by right multiplication by a non-trivial path, and no verification is given that the resulting block matrix still commutes with the objects G(M),G(N). The composition check in the proof does not imply this. If this missing verification fails, then an isomorphism G(M)≃G(N) in U need not come from M≃N, and the conclusion of Theorem 3.3.6 does not follow. The theorem's proof also relies on the same proposition to assert that G(M) is indecomposable when M is, so the gap is central. I did not find another issue of comparable weight: Proposition 1.4.1 contains a genuine error (the forgetful functor from graded to ungraded modules is faithful but not full), but it concerns Auslander–Reiten triangles and is not used in the completeness argument. The overlap with [Chr22] is a novelty caveat, not a correctness issue. Since the concern is a serious proof gap rather than a demonstrated counterexample, the appropriate verdict remains the reader's CONDITIONAL, pending a correct proof of 3.3.5(b).","tokens_in":28362,"tokens_out":33902,"duration_ms":317015,"concrete_test":"Check the missing commutation identity for the functor H in Proposition 3.3.5(b): for every pair M,N of objects in CM_rad and every T ∈ Hom_U(G(M),G(N)) satisfying T G(M)=G(N)T, verify that H(T)G(M)=G(N)H(T). This is a finite linear-algebra check in small examples. Concretely, for the gentle algebra A = k(1⇉2) with two parallel arrows a,a′, take M=(P1⊕P2, φ_a), N=(P1⊕P2, φ_{a′}) and enumerate all morphisms T between G(M) and G(N) in U; compute H(T)G(M)−G(N)H(T). If this is nonzero for some T, Proposition 3.3.5(b) is false; if it is always zero, the lift may be true but the proof still needs the omitted verification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.3.6 reduces classification to the Bondarenko–Drozd theorem via the functor G: CM_rad(A⋉) → S(Y,σ,k). The decisive step is Proposition 3.3.5(b): an isomorphism G(M)≃G(N) in the full subcategory U of objects in the image of G must lift to an isomorphism M≃N. The proof defines H on morphisms by H(T)^v_u = T^v_u if v=uα and 0 otherwise, asserts without proof that H(T) is a morphism of Im(G), verifies only that H respects composition, and concludes that isomorphisms lift. The missing check is that H(T) actually lies in Hom_S(G(M),G(N)): one must verify H(T)G(M)=G(N)H(T), the σ-condition, and triangularity. Triangularity is preserved, but commutation is not automatic; the discarded blocks T^v_u with v not of the form uα can be the only terms cancelling a contribution on one side of the equation. The proof does not use the relation T G(M)=G(N)T to establish this. Since G is not shown to be full, the assertion that H(T) is always a morphism of Im(G) is exactly the missing conservativity statement. This is load-bearing: Theorem 3.3.6 uses 3.3.5(b) both to conclude M≃M_{κ(y)} from G(M)≃G(M_{κ(y)}) and to deduce that G(M) is indecomposable when M is. Without a valid lift, Bondarenko–Drozd classifies only the matrices G(M), not the maximal Cohen–Macaulay modules themselves.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a description of the 1-periodic derived category of a finite-dimensional algebra A of finite global dimension by identifying it, via Keller and Buchweitz, with the stable category of maximal Cohen-Macaulay modules over the trivial extension A⋉. For a gentle algebra A, the paper associates to each string or band on the Opper–Plamondon–Schroll marked surface a maximal Cohen-Macaulay A⋉-module, and Theorem 3.3.6 claims that these are exactly the indecomposable objects, up to the Keller–Buchweitz equivalence. The proof follows the strategy of Bekkert–Merklen and uses the Bondarenko–Drozd classification of representations of a linearly ordered set with involution through a comparison functor G from CMrad(A⋉) to S(Y,σ,k).","tokens_in":28692,"tokens_out":19409,"duration_ms":216901,"significance":"If the main theorem were fully established, it would give a useful and explicit geometric classification of indecomposable objects in the 1-periodic derived category of a gentle algebra, complementing Christ's Fukaya-categorical construction and clarifying the role of all homotopy classes of closed curves. The paper has clear strengths: the explicit construction of string and band objects, the computation of the tλ-action and its relation to winding numbers in Proposition 2.4.2, and the overall strategy of reducing the classification to the Bondarenko–Drozd matrix problem. However, the completeness proof contains serious gaps, and the main classification is therefore not established as written.","major_comments":[{"comment":"The proof of Proposition 3.3.5(b) does not verify that the proposed morphism H(T) is actually a morphism in S(Y,σ,k). By Definition 3.1.3 one must check H(T)G(M)=G(N)H(T), the σ-condition, and triangularity. Discarding all blocks outside v=uα can break the commutation relation, and the proof never uses the equality TG(M)=G(N)T to establish this. Moreover, if α is understood to range over non-trivial paths as in the definition of G, then H(id_{G(M)}) has zero diagonal blocks and is not the identity, so H is not even a functor; if trivial α are allowed, the missing commutation check remains. Since Theorem 3.3.6 uses this proposition both to conclude M≃Mκ(y) from G(M)≃G(Mκ(y)) and to justify lifting isomorphisms, the completeness of the classification is not proved.","section":"Section 3.3, Proposition 3.3.5(b) and proof of Theorem 3.3.6"},{"comment":"The theorem proof asserts without justification that if M is indecomposable then G(M) is indecomposable in S(Y,σ,k). An additive functor need not reflect direct-sum decompositions, and Proposition 3.3.5(b) concerns only isomorphisms, not decomposability. Relatedly, Proposition 3.3.5(a) is not proved as stated: the proof concludes only that φ=0 and hence M∈projA, not M∈projA⋉. Under the identification of Proposition 1.2.10, projective A⋉-modules correspond to pairs (Q⊕Q, [[0,id],[0,0]]), not to pairs of the form (P,0). Thus the object (P,0) with P a nonzero projective A-module lies in CMrad(A⋉), is not projective as an A⋉-module, and satisfies G(M)=0. The nonvanishing of G on non-projective indecomposables, which is needed to apply the Bondarenko–Drozd classification, is therefore missing; zero-differential string objects appear to need separate treatment.","section":"Section 3.3, Theorem 3.3.6, first paragraph; Proposition 3.3.5(a)"},{"comment":"The proof of Proposition 1.4.1 claims that the forgetful functor modZ A⋉→mod A⋉ is fully faithful, but it is only faithful: after forgetting the grading there are generally additional A⋉-linear maps that are not grading-preserving. Lemma 1.4.2 requires full faithfulness to transfer almost split sequences, so the argument for Proposition 1.4.1 is invalid as written. This result is not used in the classification theorem, but it is a stated theorem of the paper.","section":"Section 1.4, Proposition 1.4.1"}],"minor_comments":[{"comment":"There are several broken cross-references: 'Corollary 1.3.12' is cited in the proofs of Theorem 1.3.5, Proposition 2.2.9 and Proposition 2.3.7, but no such corollary exists (Corollary 1.3.10 seems intended); Lemma 3.3.9 refers to 'Proposition 2.2.4', and the proof of Theorem 3.3.6 refers to 'Lemma 3.2.9', 'Lemma 3.2.10' and 'Proposition 3.2.5', which should be Proposition 3.2.4 and the corresponding items in Section 3.3.","section":"Throughout"},{"comment":"The abstract contains the typo 'indecompoable' for 'indecomposable', and Definition 3.1.1 contains 'reps.' for 'resp.'.","section":"Abstract"},{"comment":"The notation Ba(Y) is used both for a set of paths and for the set of equivalence classes under s and r; this can be confusing, although the intended meaning is usually clear from context.","section":"Section 3.1, Definitions 3.1.5-3.1.6"}],"recommendation":"major_revision","confidential_remarks":"The paper is not ready in its current form. The central classification proof relies on a conservativity statement for the comparison functor G that is not proved, and there is a further gap concerning indecomposability reflection and zero-differential objects. These issues are likely repairable within the paper's scope: one could try to prove the missing commutation for H(T), add a separate argument for objects with zero differential, or modify G so that it is full and conservative on the relevant subcategory. The Section 1.4 issue is less central but should be fixed as well. I recommend major revision rather than rejection, because the overall strategy is coherent and the intended theorem is plausible, but the current proof is incomplete."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper gives an algebraic proof of the classification of indecomposables in the 1-periodic derived category of a gentle algebra, built on Keller's orbit category equivalence, Buchweitz's singularity equivalence, and the Bondarenko–Drozd matrix problem. The headline theorem is not new—Christ already described the same indecomposables using curves in this category, and the authors say so in the introduction—but the proof is genuinely different: it goes through MCM modules over A⋉ and gives explicit module-theoretic string and band objects. That translation is clean and the BD82 machinery is applied in a natural way. The paper is honest about prior work and the framework is likely reusable.\n\nThe main soft spot is not the one the stress-test flags. The stress-test worries that the H(T) constructed in Proposition 3.3.5(b) might not be a morphism in S(Y,σ,k) because discarded off-diagonal blocks could be needed for commutation. That concern does not survive contact with the block structure. The functor G sends matrices that are block diagonal with respect to maximal paths: G(M)^v_u is nonzero only when v = uα, which forces u and v to lie in the same maximal path. Within a single maximal path, every upper-triangular block is of the form v = uα, so the map H simply zeros out the blocks that connect different maximal paths. The commutation condition T G(M) = G(N) T decouples blockwise over pairs of maximal paths, so zeroing cross-blocks preserves commutation. The proof as written is too terse—it never checks that H(T) is a morphism, and that should be added—but the lift is valid and the stress-test's structural objection does not hold.\n\nThere is, however, a genuine error in Proposition 1.4.1. Its proof uses the forgetful functor from graded to ungraded A⋉-modules as fully faithful, but it is only faithful. That invalidates the appeal to Lemma 1.4.2 and the proof of AR-triangle preservation as written. This section is peripheral to the main classification, so the damage is limited, but it needs fixing.\n\nWho is this for? Specialists in gentle algebras and periodic derived categories, and anyone interested in seeing the BD82 matrix problem appear as a classification tool for MCM modules. The main theorem is almost certainly correct, the proof is repairable, and the paper deserves a serious referee.","headline":"A solid algebraic proof of a classification already stated by Christ; one real proof error in a side section, and a terse but repairable completeness step that the stress-test concern does not actually break.","tokens_in":29257,"tokens_out":9471,"would_cite":true,"duration_ms":88506,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16G20","18E30","16E65"],"pacs":[],"model":"deepseek-v4-flash","headline":"The indecomposable objects of the 1-periodic derived category of a gentle algebra are classified by strings and bands on the associated marked surface, with each primitive closed curve contributing a family of band objects.","keywords":["gentle algebras","1-periodic derived category","orbit category","maximal Cohen-Macaulay modules","marked surfaces","string and band objects","matrix problem","winding numbers"],"falsifier":"Find two non-isomorphic indecomposable objects in $\\underline{CM}(A^{\\ltimes})$ whose images under the comparison functor $G$ are isomorphic in the matrix category $S(Y,\\sigma,k)$; such a pair would break the lifting step and the completeness direction of the classification. A direct search could start with a small gentle algebra, such as the two-square torus example used in the paper, by computing the $G$-matrices of all low-dimensional string and band objects and checking whether the matrix isomorphism classes are strictly coarser than the module isomorphism classes.","tokens_in":28103,"feed_emoji":"🔄","tokens_out":11744,"duration_ms":115947,"temperature":0.7,"pith_summary":"The paper proves a complete classification of the indecomposable objects of the 1-periodic derived category of a gentle algebra, the triangulated hull of the orbit category obtained by making the shift functor period one. The answer is geometric: after associating to the algebra a dissected marked surface, indecomposables are exactly the objects coming from homotopy classes of curves joining two marked points (strings) and, for each primitive closed curve, a family of objects indexed by indecomposable modules over the Laurent polynomial ring $k[x,x^{-1}]$ (bands). This matters because in the ordinary derived category only closed curves with zero winding number give band objects, whereas here every primitive closed curve contributes. The proof works by identifying the 1-periodic derived category with the stable category of maximal Cohen-Macaulay modules over a Gorenstein trivial-extension algebra, then using a matrix-problem classification to show that no other indecomposables exist.","feed_headline":"Gentle algebra indecomposables are exactly surface strings and bands","feed_subtitle":"Unlike the usual derived category, even non-gradable closed curves contribute band objects.","key_machinery":"The argument is carried by a chain of equivalences and a comparison functor to a matrix problem. The 1-periodic derived category is realized as the triangulated hull of the orbit category $D^b(A)/[1]$; by a theorem on triangulated orbit categories this is equivalent to the singularity category of the trivial extension $A^{\\ltimes}=A\\otimes_k k[\\varepsilon]/\\langle\\varepsilon^2\\rangle$, a Gorenstein algebra, and a further equivalence identifies that singularity category with the stable category $\\underline{CM}(A^{\\ltimes})$ of maximal Cohen-Macaulay modules. Over a gentle algebra, objects of $\\underline{CM}(A^{\\ltimes})$ are encoded by pairs $(P,\\varphi)$ with $P$ a projective $A$-module and $\\varphi^2=0$, that is, by 1-periodic complexes of projectives. The geometric input is the dissected marked surface $(S,M,\\Delta)$ attached to the algebra, whose strings and bands are in bijection with homotopy strings and homotopy bands of the algebra. The completeness input is a matrix problem: indecomposable block matrices for a linearly ordered set with involution are known to be string matrices and band matrices, with band matrices indexed additionally by indecomposable $k[x,x^{-1}]$-modules. A functor $G$ sends each object of $\\underline{CM}(A^{\\ltimes})$ to such a block matrix, and the surface string and band objects are shown to correspond exactly to the string and band matrices that occur in the image of $G$.","core_discovery":"The central claim is Theorem 3.3.6: the indecomposable objects of the stable category of maximal Cohen-Macaulay modules over $A^{\\ltimes}$ are precisely the string objects $M_y$, one for each string $y$ of the dissected marked surface $(S,M,\\Delta)$, and the band objects $M_{(y,J)}$, one for each band $y$ (a primitive closed curve) together with an indecomposable $k[x,x^{-1}]$-module $(k^n,J)$. Through the triangulated equivalence with the singularity category of $A^{\\ltimes}$, this is equivalently a complete description of indecomposables of the 1-periodic derived category $(D^b(A)/[1])^\\Delta$. A notable feature is that no grading condition is imposed on the closed curves: unlike band objects in $D^b(A)$, which require winding number zero, all primitive closed curves give families of indecomposables in the 1-periodic category.","pith_inferences":["Because the reduction to Cohen-Macaulay modules over the trivial extension works for any finite-dimensional algebra of finite global dimension, the strategy could in principle be repeated for other classes of algebras; the gentle hypothesis enters only in the final matrix-theoretic completeness step.","The paper does not give a full description of when two band parameters (orientation, starting point, module) define isomorphic objects; one could test whether the rotation and symmetry orbits it defines exhaust the equivalence relation or whether further identifications occur.","The same orbit-category construction with a different period might yield analogous curve-on-surface classifications for $m$-periodic derived categories, with band objects indexed by modules over a different Laurent ring; this is not pursued in the paper.","The $t_\\lambda$ action encoding winding numbers suggests that the 1-periodic derived category could serve as a geometric home for stability or invariant computations that need all closed curves, such as refinements of complete derived invariants for gentle algebras."],"forward_implications":["Every primitive closed curve on the surface gives rise to a family of indecomposable band objects, so the 1-periodic derived category is the setting in which all closed curves, not only gradable ones, are visible.","The winding number of a band can be recovered from the action of the $k^\\times$-automorphisms $t_\\lambda$ on the band object, since $t_\\lambda$ scales the module parameter $J$ by a power of $\\lambda$ determined by the winding number.","The triangulated functor from the ordinary bounded derived category to the 1-periodic derived category preserves Auslander-Reiten triangles, so the two categories share the same AR translation structure on the objects that survive.","The matrix-problem classification used in the proof leaves no further indecomposables, so the string and band families are exhaustive, not just a convenient geometric description."],"supporting_citations":[{"why":"Supplies the triangulated orbit category and the equivalence between the triangulated hull of $D^b(A)/[1]$ and the singularity category of $A^{\\ltimes}$.","marker":"[Kel05]"},{"why":"Supplies the equivalence between the singularity category of a Gorenstein algebra and its stable category of maximal Cohen-Macaulay modules.","marker":"[Buc21]"},{"why":"Supplies the dissected marked surface model and the bijections between homotopy strings and bands of a gentle algebra and curves on the surface.","marker":"[OPS18]"},{"why":"Supplies the classification of indecomposable objects in the matrix category $S(Y,\\sigma,k)$ as string and band matrices.","marker":"[BD82]"},{"why":"Supplies the overall completeness strategy, the definition of the linearly ordered set with involution, and the homotopy string and band combinatorics used to match matrices with surface curves.","marker":"[BM03]"},{"why":"Establishes that $A^{\\ltimes}$ is Gorenstein and that $\\underline{CM}(A^{\\ltimes})$ has Auslander-Reiten sequences, needed for the triangulated structure and AR-triangle preservation.","marker":"[AR91]"},{"why":"Provides the gradability criterion via finiteness of the $t_\\lambda$-twist orbit, used to show the grading-forgetting functor preserves almost split sequences.","marker":"[AO14]"},{"why":"Provides the combinatorial definition of winding number used to connect the $k[\\varepsilon]/\\varepsilon^2$ action on band objects to surface geometry.","marker":"[APS23]"}],"fun_headline_variants":["All primitive closed curves become bands in 1-periodic gentle algebras","No grading needed: closed curves give bands in 1-periodic gentle category","All primitive closed curves yield band objects in 1-periodic category","Gentle algebra 1-periodic indecomposables: strings and bands from all curves","1-periodic gentle category: every closed curve gives a band object"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification relies on the claim that two maximal Cohen-Macaulay modules whose associated block matrices are isomorphic in the matrix category are themselves isomorphic as modules; the proof of this lifting step is the load-bearing point of the completeness argument.","fun_headline_variants_meta":{"raw":{"variants":["All primitive closed curves become bands in 1-periodic gentle algebras","No grading needed: closed curves give bands in 1-periodic gentle category","All primitive closed curves yield band objects in 1-periodic category","Gentle algebra 1-periodic indecomposables: strings and bands from all curves","1-periodic gentle category: every closed curve gives a band object"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000765,"raw_usage":{"total_tokens":3356,"prompt_tokens":868,"completion_tokens":2488,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":484,"completion_tokens_details":{"reasoning_tokens":2390}},"tokens_in":484,"tokens_out":2488,"duration_ms":19737,"temperature":1.0,"reasoning_tokens":2390,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:51:34.108745+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find two non-isomorphic indecomposable objects in $\\underline{CM}(A^{\\ltimes})$ whose images under the comparison functor $G$ are isomorphic in the matrix category $S(Y,\\sigma,k)$; such a pair would break the lifting step and the completeness direction of the classification. A direct search could start with a small gentle algebra, such as the two-square torus example used in the paper, by computing the $G$-matrices of all low-dimensional string and band objects and checking whether the matrix isomorphism classes are strictly coarser than the module isomorphism classes.","supporting_citations":[],"review_version":1}