{"id":"5e9f7345-b393-4dfc-87d3-a5a08bb660c2","arxiv_id":"2506.19432","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"QuiverTools implements, in SageMath and Julia, algorithms for canonical decompositions, stability, Harder-Narasimhan types, Teleman quantization bounds, and Chow ring computations for quiver moduli spaces.","lead":"QuiverTools is a new open-source software package for studying the representation theory of quivers, available for both SageMath and Julia. It computes structural invariants of quiver moduli spaces, such as canonical decompositions, stability checks, Harder-Narasimhan types, and intersection-theoretic numbers.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 7 output shows a duplicate Chow ring generator symbol (x1_1bar) and omits c1(U0); if real, this is a software bug; if a typo, the central example is unverified.","rationale":"The first six sections' examples are plausible: the general subdimension vector list for the 4-Kronecker quiver matches Schofield's criterion, the stability outputs are consistent with the canonical stability parameter, and the Teleman table has no obvious inconsistency. The single anomalous output is in Section 7, where the printed generator list of the Chow ring duplicates x1_1bar and omits the generator corresponding to c1(U0). Because the paper's central claim is that QuiverTools reliably computes these invariants, an internally inconsistent displayed output is a concrete, checkable weak point that the reader's general concern about missing tests failed to pinpoint. The proposed test would settle whether the discrepancy is a software bug or a manuscript typo. Given the absence of a test suite and this concrete discrepancy, a conditional accept is appropriate: the authors should correct the example or confirm that the software output matches the intended five distinct generators before the demonstration is accepted as valid.","tokens_in":10415,"tokens_out":6088,"duration_ms":63964,"concrete_test":"Install QuiverTools v1.1 in SageMath, run the Section 7 example verbatim, and print M.chow_ring(chi=(-1,1)).gens(). Check whether the output is (x0_1bar, x0_2bar, x1_1bar, x1_2bar, x1_3bar), which is correct, or (x1_1bar, x0_2bar, x1_1bar, x1_2bar, x1_3bar), which is a bug. Also recompute the first Hilbert-series coefficient M.integral(H**0) and compare to 1; if the ring has duplicate generators, these numbers cannot be trusted.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 7 is the only place where Chow ring computations are demonstrated, and the displayed output is internally inconsistent. For d=(2,3) on the 4-Kronecker quiver, the universal family has summands U0 and U1 of ranks 2 and 3, so the Chow ring presentation of [3,7] should have five distinct generators c1(U0), c2(U0), c1(U1), c2(U1), c3(U1). The printed CH.gens() is (x1_1bar, x0_2bar, x1_1bar, x1_2bar, x1_3bar): x1_1bar appears twice, and no symbol for c1(U0), e.g. x0_1bar, appears. If this is the actual software output, the implementation of the Chow ring generators is wrong, so the subsequent Hilbert-series and anticanonical-degree computations are built on an erroneous ring presentation. If it is a transcription error, the manuscript's flagship example has not been proofread, undermining the paper's reliance on displayed outputs as evidence. This is exactly the kind of unchecked output the reader's weakest_assumption flags, but it is now concrete.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces QuiverTools, a software package available for both SageMath and Julia, aimed at computations in quiver representation theory and geometry of quiver moduli spaces. After recalling the Euler form, Kac's root-system theorem, and Schofield's theory of general subdimension vectors, the paper describes implemented algorithms for canonical decompositions, existence of (semi)stable representations, Harder-Narasimhan type enumeration, Betti numbers, Teleman quantization weights, universal-bundle weights, and intersection-theoretic invariants such as Chow ring presentations, Hilbert series, and anticanonical degrees. The running example is the 4-Kronecker quiver with dimension vector (2,3), for which a sequence of SageMath sessions is displayed. The manuscript is primarily a software announcement with mathematical background, not a proof-oriented research article.","tokens_in":10625,"tokens_out":7804,"duration_ms":88118,"significance":"If the package is correct, it would be a valuable and fairly broad computational resource: it implements recursive Schofield criteria, canonical decompositions, Harder-Narasimhan stratifications, Teleman weights from the authors' earlier papers [1,23], and Chow ring computations based on [3,7]. Strengths include the dual SageMath and Julia implementations, the versioned DOI and public repository, and the use of published theorems as the algorithmic basis. However, the paper contains no test suite, no independent cross-validation against other software, and one displayed output that is internally inconsistent with the theory it claims to implement. Because the paper's central claim is the reliability of the software, that claim is not yet independently established.","major_comments":[{"comment":"Theorem 3.7 as stated is not correct. It asserts a unique collection of dimension vectors {d_i} summing to d with ext(d_i,d_j)=0 for all i≠j, but the displayed example Q2 = ThreeVertexQuiver(1,1,1), d=(1,2,1), gives is_root((1,2,1))=True, is_schur_root((1,2,1))=False, and canonical_decomposition((1,2,1))=((0,1,0),(1,1,1)). The singleton collection {d} satisfies the pairwise condition vacuously, so uniqueness already fails unless one additionally requires that each d_i is a Schur root (equivalently ext(d_i,d_i)=0). The paragraph after the theorem states that the d_i are Schur roots, but that condition must be part of the theorem's hypothesis or conclusion for the stated characterization to be valid.","section":"Section 3, Theorem 3.7"},{"comment":"The displayed Chow ring generator tuple is internally inconsistent. For d=(2,3), the universal family has summands U0 and U1 of ranks 2 and 3, so the tautological presentation should have five distinct generators corresponding to c1(U0), c2(U0), c1(U1), c2(U1), c3(U1). The printed output (x1_1bar, x0_2bar, x1_1bar, x1_2bar, x1_3bar) contains x1_1bar twice and no x0_1bar for c1(U0). If this is actual software output, the Chow ring presentation is wrong and the subsequent Hilbert series and anticanonical degree computations are built on an incorrect ring. If it is a transcription error, the paper's flagship example has not been proofread. In either case, the output must be corrected and independently verified before the intersection-theory claims can be accepted.","section":"Section 7, CH.gens() output"},{"comment":"The paper provides no test suite, formal verification, or comparison against independent software, despite relying on nontrivial recursive algorithms in Sections 3, 5, and 6. In light of the Section 7 discrepancy, the displayed outputs cannot by themselves certify the package. Please add regression tests for at least the canonical decomposition and Schofield criterion (Section 3), Harder-Narasimhan enumeration and Betti numbers (Section 5), Teleman weights (Section 6), and the Chow ring presentation (Section 7), ideally checked against the published formulas in [1,3,7,29]. The manuscript should also state whether the displayed outputs are machine-generated by the released version v1.1 or hand-transcribed.","section":"General verification"}],"minor_comments":[{"comment":"There is a typo: 'rescallings' should be 'rescalings'.","section":"Section 4"},{"comment":"The sentence 'We illustrate this using in QuiverTools using our running example' contains a duplicated 'using'; it should be 'We illustrate this using QuiverTools on our running example'.","section":"Section 6"},{"comment":"The parameter proper=True in all_harder_narasimhan_types is used but never explained; state what it selects.","section":"Section 5"},{"comment":"The object H = M.chern_character_line_bundle(theta) is not clearly defined: it is called a 'chern character line bundle' but is then exponentiated and integrated as though it were a divisor class. Specify whether H is the Chern character, the first Chern class, or the line bundle itself.","section":"Section 7"},{"comment":"The notation Matt(α),s(α) in the definition of R(Q,d) is undefined and should be written as Mat_{t(alpha),s(alpha)} with an explanation that this is the space of matrices with rows indexed by the target and columns by the source.","section":"Equation (4)"}],"recommendation":"major_revision","confidential_remarks":"The Section 7 output inconsistency is the main blocker. If it is a transcription error, the fix is local; if it reflects a real bug in the Chow ring code, the paper's central claim about intersection theory would need substantial reworking. I would therefore not reject outright, but I would ask the authors to rerun all displayed examples from the released version, add a minimal test suite, and correct Theorem 3.7. The paper fits the scope of a software-oriented algebraic geometry venue, and the authors' prior work makes the mathematical background credible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is a software presentation for QuiverTools, a SageMath/Julia package for quiver moduli computations. What is actually new is engineering, not mathematics: the algorithms come from Schofield, Kac, Reineke, and the authors' own published formulas for Teleman weights. That is fine for a software paper, and the package looks genuinely useful: canonical decompositions, HN types, Teleman bounds, and Chow ring invariants are all things people compute by hand or ad hoc scripts today.\n\nThe exposition is clear. The mathematical statements are accurate, and the authors are honest about what the package does not do (QPA for representation-level, CoulombHiggs/msinvar for enumerative invariants). The examples are readable and the installation is documented. Credit where due: this is a real piece of work.\n\nNow the soft spots. There is no test suite, no comparison with independent software, and no formal verification. For a software paper that is common, but it means the displayed outputs are the only evidence the package is correct. That makes the following concrete: in Section 7, the only Chow ring demonstration, CH.gens() prints five symbols but lists x1_1bar twice and no generator for c1(U0). For d=(2,3) on the 4-Kronecker quiver, the universal family has summands of ranks 2 and 3, so the presentation should have five distinct generators. Either the implementation produces the wrong ring presentation, or the example was transcribed without being checked. Both possibilities undercut the paper's reliance on its examples. If it's a typo, it's the kind of typo a referee should catch before publication; if it's real, the Chow ring feature is buggy. Either way, the paper as it stands does not establish that this feature works.\n\nA second, smaller issue: the paper cites its own prior formulas for Teleman weights. That is legitimate - they are published and independently derivable - so I do not count it against them.\n\nBottom line: the core algorithms are sound and the package is likely to be a useful resource. The paper deserves peer review, not desk rejection, but it needs a minor revision: fix or explain the Section 7 output, and ideally add a shipped test suite or cross-check to back the correctness claims. I would not cite it in its current state.\n\nBest","headline":"Useful software paper with sound math, but the only displayed Chow ring computation is internally inconsistent; fix before publication.","tokens_in":11076,"tokens_out":3420,"would_cite":false,"duration_ms":33004,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16G20","14D20","14C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper presents QuiverTools, a SageMath and Julia package that computes the main numerical invariants of quiver moduli spaces from the quiver's Euler form.","keywords":["quiver representations","quiver moduli spaces","canonical decomposition","Harder-Narasimhan types","Teleman quantization","Chow rings","Euler form","SageMath"],"falsifier":"For a small case such as the 4-Kronecker quiver with $\\mathbf{d}=(2,3)$, an independent computation of the general extension group, say $\\operatorname{ext}((1,2),(1,1))$, could be checked against the package's reported list of general subdimension vectors; a single mismatch with a value computed by direct linear algebra would show the software is not reliable as claimed.","tokens_in":10231,"feed_emoji":"🧮","tokens_out":9308,"duration_ms":90126,"temperature":0.7,"pith_summary":"QuiverTools is a software package, released in both a SageMath and a Julia version, that claims to compute the main numerical and geometric invariants of quiver moduli spaces from a single input: a quiver and a dimension vector. The paper argues that the recursive Schofield criterion, expressed through the quiver's Euler form, is enough to drive general subdimension vectors, canonical decompositions into indecomposable summands, stability checks, Harder-Narasimhan types, Teleman quantization weights, and intersection-theoretic data. It walks through all of these on the 4-Kronecker quiver with dimension vector $(2,3)$, reporting explicit outputs. If the implementation is correct, the package makes a body of moduli-space calculations routine that previously required case-by-case symbolic work.","feed_headline":"One package computes quiver moduli invariants automatically","feed_subtitle":"For a quiver and dimension vector, the package outputs decompositions, stability, Teleman, and Chow invariants.","key_machinery":"The central object is the Euler form $\\langle \\mathbf{d}, \\mathbf{e}\\rangle = \\sum_{i} d_i e_i - \\sum_{\\alpha} d_{s(\\alpha)} e_{t(\\alpha)}$, whose symmetrisation gives the Kac form. The main algorithmic engine is Schofield's criterion: a subdimension vector $\\mathbf{e}$ of $\\mathbf{d}$ is general exactly when $\\operatorname{ext}(\\mathbf{e}, \\mathbf{d}-\\mathbf{e}) = 0$, with $\\operatorname{ext}$ itself computed through the Euler form and general subdimension data. That criterion feeds every other feature: the canonical decomposition via Kac's theorem, stability checks through inequalities on general subdimension vectors, Harder-Narasimhan type enumeration, the Teleman weight inequalities, and the Chow ring presentation.","core_discovery":"The paper presents QuiverTools as a unified computational implementation of known theoretical characterisations of quiver moduli invariants. The load-bearing inputs are Schofield's recursive criterion for general subdimension vectors, Kac's theorem on the canonical decomposition, the Harder-Narasimhan stratification of the unstable locus, and Euler-form formulas for Teleman weights derived in the authors' earlier work. The paper asserts that, on the running example of the 4-Kronecker quiver with dimension vector $(2,3)$ and stability parameter $(3,-2)$, the package correctly computes the general subdimension vectors, the canonical decomposition, the existence of stable representations, all eight Harder-Narasimhan types, the Teleman bounds for each stratum, the Betti numbers, and several Chow ring quantities such as the Hilbert series and the degree of the anticanonical bundle.","pith_inferences":["Editorial: A natural next step the paper does not state is to use QuiverTools to survey many quivers and dimension vectors, searching for new instances where the Teleman inequality holds and cohomology vanishes.","Editorial: One could extend the package's reliability story by generating random small quivers and comparing general subdimension vectors against brute-force sampling of representations, which would act as a de facto test suite.","Editorial: The recursive Schofield criterion is stated without complexity analysis; for larger quivers the recursion may dominate runtime, so caching or memoisation strategies would be a sensible extension.","Editorial: The displayed examples suggest the software could also support teaching and exploration of the theory, letting students test conjectures about quiver moduli numerically before attempting proofs."],"forward_implications":["A user can reproduce the paper's running example and obtain the same invariants for the 4-Kronecker quiver, including the twelve displayed Betti numbers and the degree of the anticanonical bundle.","For any quiver and stability parameter, the stability check decides whether stable or semistable representations exist, with the canonical stability parameter available as a default.","The Harder-Narasimhan type enumeration yields the complete Hesselink stratification, from which Betti numbers of the moduli space follow via the Harder-Narasimhan method.","Teleman weights let users verify the inequality $\\max \\mathcal{W}(F, \\mathbf{d}^*) < \\eta_{\\mathbf{d}^*}$ needed for cohomology vanishing, covering universal bundles, endomorphism bundles, and canonical bundles.","Chow ring computations give Euler characteristics and degrees of line bundles, and combined with the vanishing results they determine numbers of global sections."],"supporting_citations":[{"why":"Supplies Schofield's recursive criterion for general subdimension vectors and the general hom/ext formulas that power the package's core algorithms.","marker":"[29]"},{"why":"Kac's theorem identifies dimension vectors of indecomposables with roots and justifies the canonical decomposition that the package computes.","marker":"[15]"},{"why":"The authors' earlier work derives the Euler-form formulas for Teleman weights and universal bundle weights implemented in Section 6.","marker":"[1]"},{"why":"Provides the Harder-Narasimhan type enumeration and the Betti number computation method used in Section 5.","marker":"[27]"},{"why":"Introduces the stability condition for quiver representations that the package checks via general subdimension vectors.","marker":"[19]"},{"why":"Gives the Chow ring presentation for fine quiver moduli on which the intersection theory computations are based.","marker":"[3]"},{"why":"Identifies the Hesselink stratification with Harder-Narasimhan types, making the stratum enumeration and Teleman weight checks effective.","marker":"[14]"},{"why":"Establishes the tautological presentation of Chow rings of fine quiver moduli used in Section 7.","marker":"[7]"}],"fun_headline_variants":["QuiverTools: automatic quiver moduli invariants in Sage and Julia","Compute quiver decompositions, stability, and Teleman invariants","Automate quiver moduli: canonical decomposition, stability, Teleman","One package for quiver stability, decomposition, and Chow invariants"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that QuiverTools reliably computes these invariants assumes the implementation contains no bugs, especially in the recursive Schofield criterion and the Teleman weight formulas that the paper does not independently verify.","fun_headline_variants_meta":{"raw":{"variants":["QuiverTools: automatic quiver moduli invariants in Sage and Julia","Compute quiver decompositions, stability, and Teleman invariants","Automate quiver moduli: canonical decomposition, stability, Teleman","One package for quiver stability, decomposition, and Chow invariants"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000822,"raw_usage":{"total_tokens":3517,"prompt_tokens":786,"completion_tokens":2731,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":402,"completion_tokens_details":{"reasoning_tokens":2653}},"tokens_in":402,"tokens_out":2731,"duration_ms":22338,"temperature":1.0,"reasoning_tokens":2653,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:05:54.237521+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a small case such as the 4-Kronecker quiver with $\\mathbf{d}=(2,3)$, an independent computation of the general extension group, say $\\operatorname{ext}((1,2),(1,1))$, could be checked against the package's reported list of general subdimension vectors; a single mismatch with a value computed by direct linear algebra would show the software is not reliable as claimed.","supporting_citations":[{"cited_title":"Infinite root systems, representations of graphs and invariant theory","cited_arxiv_id":null,"evidence_quote":"Kac's theorem identifies dimension vectors of indecomposables with roots and justifies the canonical decomposition that the package computes."},{"cited_title":"Rigidity and Schofield's partial tilting conjecture for quiver moduli","cited_arxiv_id":"2311.17003","evidence_quote":"The authors' earlier work derives the Euler-form formulas for Teleman weights and universal bundle weights implemented in Section 6."},{"cited_title":"The Harder-Narasimhan system in quantum groups and cohomology of quiver moduli","cited_arxiv_id":null,"evidence_quote":"Provides the Harder-Narasimhan type enumeration and the Betti number computation method used in Section 5."},{"cited_title":"On Chow rings of quiver moduli","cited_arxiv_id":null,"evidence_quote":"Gives the Chow ring presentation for fine quiver moduli on which the intersection theory computations are based."},{"cited_title":"Stratifications for moduli of sheaves and moduli of quiver representa- tions","cited_arxiv_id":null,"evidence_quote":"Identifies the Hesselink stratification with Harder-Narasimhan types, making the stratum enumeration and Teleman weight checks effective."},{"cited_title":"Chow rings of fine quiver moduli are tautologically presented","cited_arxiv_id":null,"evidence_quote":"Establishes the tautological presentation of Chow rings of fine quiver moduli used in Section 7."}],"review_version":1}