{"id":"1524926e-896e-4eb1-821f-3a31fa787578","arxiv_id":"2505.00072","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any gentle algebra, the BD-gentle algebra and the CM-Auslander algebra have the same representation-finiteness and derived-discreteness as the original algebra, and this remains true under repeated applications of either construction.","lead":"This paper proves that for gentle algebras, two related algebras built from them, the BD-gentle algebra and the CM-Auslander algebra, keep the same finiteness behavior and the same derived-discreteness as the original algebra. The result matters because representation theorists use these constructions to organize module categories, and knowing the representation type is preserved makes them more reliable tools.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The converse directions of Lemmas 4.3, 4.8, and 4.9 rely on an unproven structural claim about bands in BD(A) that is demonstrably false in a simple example, so the collapse property is not established.","rationale":"The reader's weakest assumption correctly identifies the structural classification of bands/homotopy bands in BD(A) and ACMA as the linchpin of Theorems 1.1 and 1.2. My analysis confirms this is the most load-bearing concern: the proofs of Lemmas 4.3, 4.8, and 4.9 do not establish the collapse property, and the specific odd-length claim in Lemma 4.3 is false, as shown by the Kronecker example. This is more than a presentation error; it means the proof of the central claim is not valid as written. However, the Kronecker example still collapses, and the geometric remarks (Remarks 4.4 and 4.10) suggest the underlying statement may be true, so a correction or a new proof of the collapse property is plausible. The reader's CONDITIONAL verdict is therefore appropriate: the paper should not be accepted until this gap is resolved. My recommendation is UNCHANGED relative to the reader's verdict. The other issues noted by the reader (the swapped cases in the displayed formula in Lemma 4.9, the 'representation-discrete' wording in Theorem 4.13, and the wrong proposition citation) are secondary presentation problems that do not affect this assessment.","tokens_in":21845,"tokens_out":42910,"duration_ms":409330,"concrete_test":"Perform a brute-force check on all gentle algebras with up to 5 vertices (enumerating quivers with indegree/outdegree at most 2 and length-2 relations), construct BD(A) and ACMA via Theorems 2.7 and 2.9, and use standard string-band enumeration to test whether existence of bands and of homotopy bands is invariant under BD and (−)CMA. If any example has a band or homotopy band in BD(A) or ACMA while A has none, the main theorems fail. Independently, re-examine the Kronecker example: verify that the even-length band (1→2'_α→2)(2→2'_β→1) is valid and collapses to a band in A, and check whether every band in BD(A) with even-length letters can be collapsed by deleting primed vertices; this tests the corrected form of Lemma 4.3 that the paper needs.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorems 4.6 and 4.13 depend on the assertions that every band (resp. homotopy band) in BD(A), and every homotopy band in ACMA, has a specific rigid form: in Lemma 4.3 the proof claims each letter of a band in BD(A) has odd length, and in Lemma 4.8 that every homotopy band is of the form (4.3). These assertions are not proved and the odd-length claim is actually false. For the Kronecker gentle algebra A = k(1⇉2), the algebra BD(A) has quiver with two parallel paths 1'_α→1→2'_α→2 and 1'_β→1→2'_β→2. The cyclic string ρ = (1→2'_α→2)(2→2'_β→1) is a genuine band in BD(A): both letters have even length 2, the endpoints are unprimed, and the two junctions use different primed twins (2'_α vs 2'_β and 2'_β vs 2'_α). This contradicts the lemma's assertion that all band letters are odd. The band does collapse to the band αβ^{-1} in A, so this particular example does not disprove the theorem; however, it shows the stated proof mechanism is invalid. The paper gives no alternative argument that deleting all primed vertices (or collapsing α±/α∓ pairs in the CM-Auslander case) always sends a band/homotopy band to one in A, nor that a band cannot use the doubled vertices in a mixed way that fails to collapse. If such a band exists, BD(A) or ACMA could be representation-infinite or non-derived-discrete while A is finite or discrete, breaking Theorems 1.1 and 1.2. The analogous step in Lemma 4.9 is asserted to be 'similar' with no details. This is the load-bearing gap in the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies two algebras attached to a gentle algebra A: the BD-gentle algebra BD(A) of Burban–Drozd and the CM-Auslander algebra A_CMA. It claims that A, BD(A), and A_CMA have the same representation-finiteness and the same derived-discreteness, and that both properties are preserved under arbitrary finite iterations of the two operations (Theorems 1.1 and 1.2). The strategy is to compare strings/bands and homotopy strings/homotopy bands of the three bound quivers: Lemmas 4.3, 4.8, and 4.9 assert correspondences of bands and homotopy bands, and these are then combined with the Butler–Ringel correspondence and the Bekkert–Merklen criterion for derived-discreteness. The CM-Auslander part is largely delegated to Chen–Lu [15]; the BD part is the new content.","tokens_in":22197,"tokens_out":8155,"duration_ms":83786,"significance":"If correct, the result is a clean invariance statement: the two geometrically natural constructions BD and (-)_CMA do not change whether a gentle algebra is representation-finite or derived-discrete, and can be iterated freely. The surface constructions in Section 3 are a useful contribution, and the overall strategy of comparing bands is natural. However, the BD part is currently supported by incomplete structural assertions about the form of bands in BD(A), one of which is false as stated; until those assertions are replaced by a valid collapse argument, the main theorems are conditional.","major_comments":[{"comment":"The converse direction of Lemma 4.3 is not proved. The proof assumes that every band of BD(A) has the displayed form (4.1), with each letter q_k of odd length and arising from a permitted path in A by inserting primed vertices. This is false: for the Kronecker gentle algebra A = k(1=>2) with arrows α and β, the quiver of BD(A) contains the paths 1'_α -> 1 -> 2'_α -> 2 and 1'_β -> 1 -> 2'_β -> 2, and the cyclic string ρ = (1 -> 2'_α -> 2)(2 -> 2'_β -> 1) is a band in BD(A) whose two letters have length 2. This contradicts the assertion that all letters are odd. The example does collapse to the band αβ^{-1} in A, so it does not disprove the theorem, but it shows the given proof mechanism is invalid. The paper needs a separate argument that deleting all primed vertices sends every band of BD(A) to a band of A, or that bands of BD(A) cannot mix the doubled vertices in other ways.","section":"4.1, Lemma 4.3"},{"comment":"The converse direction of Lemma 4.8 is asserted rather than proved: after constructing a homotopy band from A to BD(A), the proof says that a homotopy band in BD(A) 'must be of the form given by (4.3)' and then removes primed vertices. No argument is given that excludes homotopy bands whose letters mix different doubling scales or use the primed vertices in a non-alternating way. Since Proposition 4.12 and Theorem 1.2 depend on this converse, this is a load-bearing gap.","section":"4.2, Lemma 4.8"},{"comment":"Lemma 4.9 also has a load-bearing gap and a formula error. In the forward direction, the displayed definition of y_{i,j} is the reverse of the preceding sentence: an arrow on a forbidden cycle should become the two-letter path x^-_{i,j} x^+_{i,j}, not remain x_{i,j}. In the converse, the claim that α+ and α- must occur together in every homotopy letter is not justified; α+ can start a directed letter at the vertex αA without a preceding α- in the same letter. The equivalence of homotopy bands between A and A_CMA is therefore not rigorously established.","section":"4.2, Lemma 4.9"}],"minor_comments":[{"comment":"The statements say 'representation-discrete', but the property defined and proved in Section 4.2 is 'derived-discrete'; the statements should be corrected for consistency.","section":"Theorems 1.2 and 4.13"},{"comment":"The two cases in the displayed formula for y_{i,j} are swapped relative to the surrounding text; one of the two needs to be corrected.","section":"4.2, Lemma 4.9"},{"comment":"There are several typos, e.g., 't he' in the Abstract, 'Baurban–Drozd' in the title of Section 2.2, and inconsistent use of 'permissible' versus 'permitted'; in Example 3.3 the reference to 'Figure 3.3, III' appears to mean Figure 3.4, III.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The BD part is the paper's main new claim, and its proof is incomplete as written. The Kronecker counterexample to the proof mechanism still collapses to an ordinary band in A, so the theorem itself may well be salvageable, but the authors need to replace the asserted 'must be of the form' claims with a genuine structural argument. The final version should also make explicit that the CM-Auslander representation-type part is essentially Chen–Lu's theorem and focus the new arguments on the BD construction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this paper's main theorem is probably true, but the proof of the BD half has a hole that the authors don't appear to see. The result is that for a gentle algebra A, the BD-gentle algebra BD(A) and the CM-Auslander algebra A_CMA have the same representation-finiteness and derived-discreteness as A. The CM-Auslander half is already in Chen–Lu; the BD half is new. The paper gives a nice surface-based picture for both constructions and reduces everything to the existence of bands/homotopy bands. That's the right approach.\n\nThe problem is in the converses of Lemmas 4.3 and 4.8. The proof claims that every band (resp. homotopy band) in BD(A) must have letters of a very specific alternating form, with odd length. That is false. Take A = k(1⇉2), the Kronecker. BD(A) has a band 1→2'_α→2→2'_β→1 (with the obvious naming), whose two letters both have length 2. So the 'odd length' claim is not a minor gap; it is simply wrong. The band does collapse to αβ^{-1} in A, so the theorem survives this example, but the proof gives no reason to think every band collapses. If there is a band in BD(A) whose primed vertices don't delete to a band in A, the whole equivalence breaks. The paper offers no alternative argument. That is load-bearing.\n\nThere are also the usual fixable typos: Lemma 4.9's displayed formula has the two cases swapped relative to the prose, Theorem 4.13 says 'representation-discrete' where it means 'derived-discrete', and its proof cites Proposition 4.5 instead of 4.12.\n\nOn the positive side, the surface constructions in Section 3 are genuinely useful, the reduction to bands is the right high-level strategy, and the iterated statement is a harmless corollary once the one-step case is fixed. The authors are not sloppy in their overall thinking; the gap is specific and local.\n\nWho is this for? People working on gentle algebras, BD algebras, or derived-discreteness. If you are in that group, you should know the result and treat it as a conjecture with a good sketch, not a proven theorem.\n\nRecommendation: send it to a serious referee, but with a clear note that the proof of Lemmas 4.3 and 4.8 needs to be reworked. A desk reject would be too harsh; accepting as-is would be wrong. I'd want to see a revised version.","headline":"The BD half of the main theorem is new and plausible, but the key lemma's converse relies on a false structural claim about bands in the doubled quiver; the paper needs a real proof repair, not just typo fixes.","tokens_in":22779,"tokens_out":13728,"would_cite":false,"duration_ms":123184,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16G10","16G60"],"pacs":[],"model":"deepseek-v4-flash","headline":"Applying BD or CM-Auslander constructions to a gentle algebra preserves its representation type and derived-discreteness.","keywords":["gentle algebras","representation-finite","derived-discrete","BD-gentle algebra","CM-Auslander algebra","bands","homotopy bands","marked surfaces"],"falsifier":"Take a gentle algebra $A$, compute the bound quiver of $BD(A)$, and enumerate all cyclic strings on it; if any band has a letter whose length is not odd or whose primed vertices do not occur exactly once per original arrow, then deleting primed vertices does not collapse it to a band of $A$. Exhibiting such a band for an $A$ that itself has no bands would give a representation-infinite $BD(A)$ with representation-finite $A$, directly contradicting Theorem 1.1.","tokens_in":21592,"feed_emoji":"🔁","tokens_out":10861,"duration_ms":104263,"temperature":0.7,"pith_summary":"Gentle algebras are a broadly studied class of finite-dimensional algebras governed by a small set of combinatorial rules. The paper proves that two standard ways of manufacturing a new gentle algebra from an old one—the BD-gentle algebra and the CM-Auslander algebra—preserve the two basic representation-type dichotomies: representation-finiteness and derived-discreteness. A gentle algebra is representation-finite exactly when its BD-gentle and CM-Auslander algebras are, and likewise for derived-discreteness. Because both constructions output gentle algebras, the equivalence survives any finite sequence of the two operations, so the class of representation-finite (or derived-discrete) gentle algebras is closed under repeatedly applying $BD$ and $(-)^{CMA}$.","feed_headline":"BD and CM-Auslander constructions preserve representation type","feed_subtitle":"For gentle algebras, being representation-finite or derived-discrete survives repeated application of both operations.","key_machinery":"The load-bearing objects are bands and homotopy bands. A band is a cyclic string, i.e. a word in arrows and formal inverses whose square is again a string and that is not a nontrivial power; one band forces infinitely many indecomposable modules. A homotopy band is the derived-category analogue, with equal numbers of direct and inverse letters, and its absence exactly characterizes derived-discreteness for gentle algebras. The technical core is a letter-by-letter translation: in $BD(A)$ every permitted path of $A$ is doubled by inserting new primed vertices, while in $A_{CMA}$ each arrow lying on a forbidden cycle splits into a pair of arrows. The lemmas show these translations send bands to bands and homotopy bands to homotopy bands, and that deleting the primed vertices or collapsing the split arrows recovers a band of $A$.","core_discovery":"Let $A$ be gentle, let $B=BD(A)$ be its BD-gentle algebra, and let $C=A_{CMA}$ be its CM-Auslander algebra. The paper claims that $A$, $B$, and $C$ are simultaneously representation-finite, and simultaneously derived-discrete; moreover, for any finite sequence of the two operations $BD$ and $(-)^{CMA}$, the resulting algebra has the same two properties as $A$. The proof works by tracking bands and homotopy bands. A gentle algebra is representation-infinite exactly when its bound quiver contains a band, and it fails to be derived-discrete exactly when a homotopy band exists; the paper constructs explicit path-level translations showing that bands and homotopy bands occur in $BD(A)$ or $A_{CMA}$ if and only if they occur in $A$. Thus the absence of these cyclic structures, which is what finiteness and discreteness mean for gentle algebras, is inherited in both directions.","pith_inferences":["The band-collapsing correspondence suggests a stronger structural claim the paper does not make: the set of bands up to rotation may be preserved bijectively, not just its emptiness, which would imply finer invariants like the number of one-parameter module families are also unchanged.","The marked-surface proofs in the remarks point to a testable extension: $BD$ and CM-Auslander may induce homotopy equivalences between the relevant curve spaces in the geometric models, which would let band and homotopy-band correspondences be upgraded to functorial statements.","One could test whether the same preservation extends to neighbouring dichotomies such as domesticity or $\\tau$-tilting finiteness; the paper's band-counting method alone does not settle those, and they would require additional arguments."],"forward_implications":["For any gentle algebra $A$, representation-finiteness of $BD(A)$ or $A_{CMA}$ is equivalent to representation-finiteness of $A$, so neither construction can create or destroy the finite/infinite dichotomy.","Derived-discreteness is likewise invariant: repeated application of $BD$ and $(-)^{CMA}$ does not create or destroy infinite families of indecomposable objects in the bounded derived category.","The class of gentle algebras that are representation-finite or derived-discrete is closed under the semigroup generated by $BD$ and $(-)^{CMA}$, so any finite composition preserves the property.","For checking these properties, it suffices to inspect $A$ itself: a band or homotopy band search in $A$ decides the status of every iterated image."],"supporting_citations":[{"why":"Classification of modules over string algebras by strings and bands; used to equate existence of a band with representation-infiniteness.","marker":"[10]"},{"why":"Criterion that a gentle algebra is derived-discrete exactly when its bound quiver has no homotopy band; backbone of Proposition 4.12.","marker":"[7]"},{"why":"Shows the CM-Auslander algebra of a gentle algebra is gentle and that $A$ and $A_{CMA}$ have the same representation type; starting point for the CM-Auslander part.","marker":"[15]"},{"why":"Introduces the BD-gentle algebra as a matrix algebra and supplies the structural decomposition used in Theorem 2.7 and Lemma 4.3.","marker":"[9]"},{"why":"Introduces homotopy strings and homotopy bands in the bounded derived category; the objects tracked in Lemmas 4.8 and 4.9.","marker":"[2]"},{"why":"Geometric model for graded skew-gentle algebras; provides the admissible-curve description of homotopy strings used in Remarks 4.4 and 4.10.","marker":"[26]"},{"why":"Geometric model of the module category of a gentle algebra via marked surfaces; used in the surface proof that bands are preserved under $BD$.","marker":"[6]"},{"why":"Geometric model of the derived category of gentle algebras; underlies the correspondence between homotopy strings and admissible curves.","marker":"[25]"}],"fun_headline_variants":["BD and CM-Auslander preserve gentle algebra representation type","For gentle algebras, BD and CM-Auslander leave representation type unchanged","Gentle algebra finiteness and discreteness survive BD and CM-Auslander","BD and CM-Auslander ops don't change gentle algebra representation type"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reverse direction assumes every band or homotopy band in $BD(A)$ has a rigid alternating shape—each original arrow doubled into exactly one primed and one unprimed step—so that deleting all primed vertices always collapses it to a band in $A$; the CM-Auslander argument makes an analogous structural assumption about split arrows.","fun_headline_variants_meta":{"raw":{"variants":["BD and CM-Auslander preserve gentle algebra representation type","For gentle algebras, BD and CM-Auslander leave representation type unchanged","Gentle algebra finiteness and discreteness survive BD and CM-Auslander","BD and CM-Auslander ops don't change gentle algebra representation type"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000381,"raw_usage":{"total_tokens":1945,"prompt_tokens":792,"completion_tokens":1153,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":408,"completion_tokens_details":{"reasoning_tokens":1076}},"tokens_in":408,"tokens_out":1153,"duration_ms":10536,"temperature":1.0,"reasoning_tokens":1076,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:51:56.535983+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a gentle algebra $A$, compute the bound quiver of $BD(A)$, and enumerate all cyclic strings on it; if any band has a letter whose length is not odd or whose primed vertices do not occur exactly once per original arrow, then deleting primed vertices does not collapse it to a band of $A$. Exhibiting such a band for an $A$ that itself has no bands would give a representation-infinite $BD(A)$ with representation-finite $A$, directly contradicting Theorem 1.1.","supporting_citations":[{"cited_title":"On the derived categories of gentle and skew-gentle algebras: homological algebra and matrix problems","cited_arxiv_id":"1706.08358","evidence_quote":"Introduces the BD-gentle algebra as a matrix algebra and supplies the structural decomposition used in Theorem 2.7 and Lemma 4.3."}],"review_version":1}