{"id":"ed884040-fea5-4f57-b99d-f06c8c2caab4","arxiv_id":"2507.11982","paper_version":1,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":1.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"An expository review explaining that rounding of complex log spaces generalizes real oriented blowups and builds canonical representatives for links of singularities and Milnor fibers.","lead":"This paper is a survey of how to turn algebro-geometric boundaries into canonical topological boundaries using real oriented blowups and Kato-Nakayama roundings. It explains the theory through a running example: the passage to polar coordinates.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing step is the unproved passage in Corollary 4.53 from Nakayama-Ogus local triviality (Theorem 4.49) to a Milnor tube with boundary; the text dismisses this as 'straightforward' without supplying the argument.","rationale":"The paper is an expository survey whose central assertion is that rounding canonically replaces the algebro-geometric boundary by a topological boundary and yields canonical representatives for links and Milnor fibers. The reader's verdict is UNVERDICTED because no new theorem is advanced. My stress-test focuses on the most load-bearing consequence, Corollary 4.53, which claims that the rounding of the log enhancement of a smoothing is a fibration isomorphic to the Milnor fibration. The proof is not given; the paper cites Nakayama-Ogus's Theorem 4.49 and asserts a boundary version of Corollary 4.51 as 'straightforward.' This is precisely where the central claim is least secure: Theorem 4.49 is quoted from the literature, and its extension to a Milnor tube with boundary is nontrivial and unproved. This agrees partially with the reader's weakest_assumption, which identified the reliance on external results; my concern is more specific, pointing to an internal gap in the application of those results. I do not claim the statement is false; I claim it is unverified in the text. Since the paper is a survey and no new result is offered, this gap does not change the UNVERDICTED verdict, but it does mean that the singularity-theory applications should be treated as conditional on the boundary version of the Nakayama-Ogus theorem being correct. The proposed concrete test, checking the xy example, would settle whether the claimed fiberwise homeomorphism holds in the simplest case; if it fails there, the central claim is refuted, and if it holds, it supports the survey's exposition.","tokens_in":46655,"tokens_out":25638,"duration_ms":285871,"concrete_test":"Write out the proof of Corollary 4.53 by explicitly verifying the hypotheses of Theorem 4.49 (properness, exactness, relative log smoothness) for the log enhancement of f composed with pi on a Milnor tube with boundary, and prove that the restriction of the rounding fibration to the boundary circle is fiberwise homeomorphic to the Milnor fibration. Apply this to the example f(z1,z2) = z1*z2 on C^2 with the identity resolution: if the fiber over each boundary point of the cylinder (D^dagger_0)^odot is not homeomorphic to the Milnor fiber (an annulus), then the claimed isomorphism in Corollary 4.53 is false.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim about canonical representatives of Milnor fibrations rests on Corollary 4.53, which states that the rounding of the log enhancement of f composed with pi is a locally trivial fibration whose restriction to the boundary circle is isomorphic to the Milnor fibration. The paper does not prove this corollary. Instead, it relies on Theorem 4.49, quoted from Nakayama-Ogus, which is stated for proper morphisms of log complex analytic spaces. A Milnor tube representative has a topological boundary, and the text explicitly notes the difference: 'V := X has a topological boundary. However, since f composed with pi is locally trivial near that boundary, it is straightforward to show that a version of Corollary 4.51 applies.' This is the weakest load-bearing step: the boundary version requires an additional argument (for example, a collar argument) that is not supplied, and the claimed isomorphism with the Milnor fibration over the circle depends on it. If the boundary version of local triviality fails, then the rounding fibration need not be locally trivial, and its restriction to the boundary circle need not be fiberwise homeomorphic to the Milnor fibration. The paper flags this as a gap by acknowledging the difference, but does not fill it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This expository paper develops the circle of ideas around real oriented blowups and Kato–Nakayama roundings, starting from polar coordinates and proceeding through affine and general toric varieties, toroidal varieties, A'Campo real oriented blowups, log spaces, prelog structures, charts, and the rounding functor. The final sections state Nakayama–Ogus local triviality and draw consequences for singularity theory: canonical representatives of links of isolated singularities and of Milnor fibrations of smoothings, obtained by rounding log enhancements of resolutions. The paper is explicitly a survey: all main theorems are attributed to the literature, with many proofs omitted or left as exercises.","tokens_in":46924,"tokens_out":4351,"duration_ms":54778,"significance":"If the external results it relies on are correct, the paper provides a useful and unusually pedagogical bridge from elementary polar coordinates to current log-geometric techniques in singularity theory. Its main conceptual claim — that rounding canonically replaces the algebraic boundary by a topological boundary, and that this gives canonical Milnor-fibration representatives — is significant for the intended audience. The manuscript is transparent about its sources, carefully attributes results to Kato, Kato–Nakayama, Nakayama–Ogus, Kajiwara–Nakayama, Argüz, and the author's joint work with Cueto and Stepanov, and it contains no circular derivations or fitted parameters. As a survey it does not prove the major external theorems, but that is appropriate. The main weakness is that one of its advertised applications, Corollary 4.53, contains a genuinely nontrivial boundary-with-corners step that is asserted rather than proved.","major_comments":[{"comment":"The passage from Theorem 4.49/Corollary 4.51 to the Milnor-tube setting is load-bearing and is not actually supplied. Theorem 4.49 is stated for a proper morphism of log complex analytic spaces, while the Milnor tube representative has a topological boundary and f∘π is not proper on the closed tube; the text acknowledges this difference but only says that 'since f∘π is locally trivial near that boundary, it is straightforward to show that a version of Corollary 4.51 applies.' That is exactly the step that needs a collar or relative Ehresmann argument, and it is not given. Moreover, the final assertion that the restriction to the boundary circle is 'isomorphic to the Milnor fibration' is stronger than mere local triviality: it also requires a comparison of the rounding fibers over the circle with the classical Milnor fibers over the corresponding angles, including the monodromy identification. Because Corollary 4.53 is the stated mechanism for obtaining canonical representatives of Milnor fibrations, this gap should be closed by a proof, a precise reference, or a clearly marked reduction to an existing result in [19] or [57].","section":"§4.8, Corollary 4.53"},{"comment":"The link representative (E†)^⊙ is asserted to be a canonical representative of ∂(X,x) via Proposition 4.44, but the argument depends on the identification of rounding with the real oriented blowup for toroidal varieties, which is quoted from [42, Proposition A.1] and [6, Proposition 2.4]. This is acceptable for a survey, but the text should state more explicitly that the canonicity of the link representative is only relative to a chosen snc resolution and that the invariance under different resolutions is not established here; the current wording in §1.4 and §4.9 may leave a reader with the impression that the representative is independent of all choices.","section":"§1.4 and §4.9, applications to links"}],"minor_comments":[{"comment":"In the statement of Theorem 4.49, the symbol f is used for the morphism, while the surrounding text and Corollary 4.51 use ϕ or φ; this should be unified to avoid confusion.","section":"§4.8, Theorem 4.49"},{"comment":"The word 'pullbak' should be 'pullback'.","section":"§4.4, Definition 4.25"},{"comment":"The proof of Proposition 3.9 invokes Proposition 4.43, which is stated much later and itself relies on log-geometric notions. A direct toric-chart argument, or an explicit statement that this is a forward reference in an expository ordering, would improve readability.","section":"§3.2, Proposition 3.9"},{"comment":"The authors of the classical reference [48] are spelled 'Kempf, Knudsen, Mumford and Saint-Donat', but the text twice writes 'Kempf, Knudson, Mumford and Saint-Donat'; the spelling should be corrected.","section":"§2.7"},{"comment":"In the paragraph after Definition 3.8, 'real oriented real blowup' is a typo for 'real oriented blowup'.","section":"§3.1, Definition 3.8"},{"comment":"The sentence 'We will not give the definitions of the terms involved in the statement which we have not discussed until now' is helpful, but it would be even more useful to point the reader to the precise locations in [57] where relative coherence, exactness, relative log smoothness, and verticality are defined.","section":"§4.8, after Corollary 4.51"}],"recommendation":"major_revision","confidential_remarks":"The paper is a well-written survey with a clear pedagogical arc, and the reliance on external theorems is appropriate. The main issue is that the most advertised application — Corollary 4.53 — contains a genuine topological gap that is not merely expository. I would recommend major revision with the request that the author either prove the boundary/collar version of Nakayama–Ogus local triviality, or explicitly downgrade the claim to a citation of a proved statement from [19] or another source. If the gap is filled, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid expository chapter, not a research announcement. It introduces no new theorems, but it does a real service by organizing the passage from polar coordinates through toric, toroidal, and log geometries to Kato-Nakayama rounding, with careful attributions to A'Campo, Kato-Nakayama, Nakayama-Ogus, and the author's own prior work. The writing is clear and pedagogical; the diagrams and repeated examples (the origin in C) genuinely help. The author also flags the non-standard notation W⊙ and 'fan of monoids' as terminological, so there is no overclaiming of novelty.\n\nWhat it does well: it gives a unified, coordinate-free route to real oriented blowups, explains why rounding generalizes A'Campo's operation, and states Nakayama-Ogus local triviality with consequences for Milnor fibrations. The attributions look accurate; I spot-checked several and found nothing misquoted. The typos mentioned (ϕ vs f in Theorem 4.49, 'pullbak' in Definition 4.25) are minor and easily fixed.\n\nThe soft spots are also real, though not fatal for a survey. First, the proof of Proposition 3.9 invokes Proposition 4.43, which comes later; that is a presentation choice, not a circularity, but a reader might stumble. Second and more importantly, Corollary 4.53 is the load-bearing step for the Milnor fibration claim. The text explicitly notes that the Milnor tube has a topological boundary, and then says it is 'straightforward to show' that a version of Corollary 4.51 applies. That is a genuine gap in the exposition. I believe the gap is fillable—the standard collar argument should do the job—but the paper does not supply it. Since this corollary is the payoff of the whole chapter, the authors should either prove it or explicitly cite a place where it is proved. In the current state, an otherwise careful reader is left to supply a nontrivial topological argument.\n\nWho benefits? Graduate students and researchers in singularity theory who want a working knowledge of log geometry via polar coordinates; also people seeking a map of the literature. It is not for someone hunting new theorems. I would send it to a referee: the expository level is high, the statements are mostly checkable, and the one gap I found is a fixable omission rather than a sign of error. The referee should be asked to verify Corollary 4.53 and the local triviality claim.\n\nMy own verdict: engage with it, but require the authors to fill the boundary argument before final acceptance. It deserves a serious referee, and once the gap is addressed it will be a reliable reference.","headline":"A careful, genuinely useful survey of real oriented blowups and roundings; its main soft spot is a hand-waved boundary argument in Corollary 4.53.","tokens_in":47421,"tokens_out":2121,"would_cite":false,"duration_ms":25055,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14A21","14B05","14M25","32S05","32S55"],"pacs":[],"model":"deepseek-v4-flash","headline":"The rounding of complex log spaces is a canonical real oriented blowup: it cuts any complex toroidal variety along its toroidal boundary and yields canonical representatives of links and Milnor fibers in singularity theory.","keywords":["rounding","Kato-Nakayama spaces","real oriented blowups","log structures","toroidal varieties","Milnor fibers","links of singularities","polar coordinates"],"falsifier":"Take a Brieskorn hypersurface such as $x^2 + y^3 + z^5 = 0$ with its standard toroidal resolution: compute the rounding of the restriction to the exceptional divisor $E$ of the divisorial log structure, and compare the resulting space, with its projection to $E$, to the classical link (a Brieskorn sphere). If the two are not homeomorphic as spaces over $E$, the claim that rounding gives a canonical representative of the link is false.","tokens_in":46463,"feed_emoji":"⭕","tokens_out":11105,"duration_ms":121817,"temperature":0.7,"pith_summary":"This survey argues that Kato and Nakayama's rounding operation, also called Betti realization, is the right canonical form of A'Campo's real oriented blowup. The paper builds toric, toroidal, and logarithmic geometries from the single example of the passage to polar coordinates, and then presents rounding as a functor that replaces the algebro-geometric boundary of a complex log space by a topological boundary. If the claims are correct, every complex toroidal variety has a canonical manifold-with-boundary model, and every isolated complex analytic singularity and every smoothing has canonical representatives for its link and Milnor fiber once a toroidal resolution is chosen. The practical payoff is that these representatives come with canonical projections and are canonically decomposed into pieces indexed by the components of the exceptional divisor or central fiber, so no choices of tubular neighborhoods are involved.","feed_headline":"Roundings give canonical links and Milnor fibers for singularities","feed_subtitle":"One log-space operation generalizes A'Campo's real oriented blowup and ends the need to choose tubular neighborhoods.","key_machinery":"The central object is the rounding functor $W\\mapsto W^\\odot$, defined for a complex log space $W$ by $W^\\odot := \\{(x,u) : x\\in |W|,\\ u\\in\\mathrm{Hom}(M_{W,x},S^1),\\ u(f)=\\mathrm{sign}(f(x))\\text{ for all units } f\\}$, where $(M_{W,x},\\alpha_{W,x})$ is the stalk of the log structure. Log structures—sheaves of monoids with an evaluation map to the structure sheaf—package the boundary divisor without choosing a tubular neighborhood. For the complex plane with the log structure of the origin, $W^\\odot = \\mathbb{R}_{\\geq 0}\\times S^1$ and $\\tau_W$ is $\\tau_{\\mathbb{C}}$; for a toroidal variety, the same construction gives the real oriented blowup. Applied to a log morphism satisfying properness, separatedness, exactness, and relative log smoothness, rounding becomes a locally trivial fibration by the Nakayama–Ogus theorem. The polar log point $\\Pi_{\\mathbb{C}}^{\\mathrm{pol}}$, i.e. the monoid $\\mathbb{R}_{\\geq 0}\\times S^1$ acting on $\\mathbb{C}$, serves as the universal 'rounding point': $W^\\odot$ is exactly the set of log-space morphisms from $\\Pi_{\\mathbb{C}}^{\\mathrm{pol}}$ to $W$.","core_discovery":"The central discovery presented is that rounding is a functorial generalization of the real oriented blowup and is already forced by the ordinary polar-coordinates map $\\tau_{\\mathbb{C}}: \\mathbb{R}_{\\geq 0}\\times S^1 \\to \\mathbb{C}$. For any complex log space $W$, its rounding $W^\\odot$ is the set of pairs $(x,u)$ with $x\\in |W|$ and $u$ a monoid homomorphism from the stalk of the log structure at $x$ to $S^1$ extending the sign map; the rounding map $\\tau_W: W^\\odot \\to |W|$ sends $(x,u)$ to $x$. On a toroidal variety $X$ with its canonical divisorial log structure, this $W^\\odot$ is the real oriented blowup along the toroidal boundary, a topological manifold-with-boundary whose boundary is the preimage of $\\partial X$ and is a canonical representative of the boundary of any tubular neighborhood of $\\partial X$. In singularity theory, restricting the divisorial log structure to an exceptional divisor $E$ with simple normal crossings and rounding gives a representative of the link of an isolated singularity, while rounding the log enhancement of a resolution of a smoothing yields a locally trivial fibration isomorphic to the Milnor fibration. The engine for the fibration statement is the Nakayama–Ogus local triviality theorem.","pith_inferences":["Beyond the paper: the functoriality of rounding suggests that one could define link invariants directly from log structures without fixing a resolution, provided a resolution-independent or derived version of the boundary is used.","Beyond the paper: the reinterpretation $W^\\odot = \\mathrm{Hom}(\\Pi_{\\mathbb{C}}^{\\mathrm{pol}}, W)$ is a template: replacing the polar log point by other log points yields variants such as extended roundings with canonical geometric monodromy, which could be developed into a general theory of canonical boundary replacements for degenerations.","Beyond the paper: the canonical representatives may allow comparison of Milnor fibrations across different resolutions purely at the level of log structures, potentially simplifying proofs of resolution-independence conjectures for wider singularity classes.","Beyond the paper: the toroidal version of Corollary 4.53 can be tested explicitly on Newton non-degenerate singularities by computing roundings from local tropicalizations and checking that the resulting fibration is independent of the chosen tropical subdivision."],"forward_implications":["Any complex toroidal variety $X$ comes with a canonical topological manifold-with-boundary $X^\\odot$ whose boundary is homeomorphic to the boundary of every tubular neighborhood of $\\partial X$.","Links of isolated complex analytic singularities have canonical representatives built from roundings of divisorial log structures; each representative carries a canonical projection to the exceptional divisor.","Milnor fibrations of smoothings of complex singularities are represented, up to isomorphism, by roundings of the log enhancement of $f\\circ\\pi$; the roundings are locally trivial fibrations and are canonically decomposed into pieces above the components of the central fiber.","Because rounding is functorial, any factorization of resolutions produces compatible maps between link representatives, yielding an inverse system of canonical representatives for the link.","The same machinery works when the resolution is only toroidal and not necessarily smooth, opening singularities to the use of tropical and toroidal modifications."],"supporting_citations":[{"why":"Supplies the Nakayama–Ogus local triviality theorem (Theorem 4.49), from which the Milnor-fibration representative follows.","marker":"[57]"},{"why":"Introduces the rounding functor and its basic properties; the survey's central object.","marker":"[46]"},{"why":"Quoted for Proposition 4.43, identifying roundings of toric varieties with their real oriented blowups.","marker":"[42]"},{"why":"Also quoted for Proposition 4.43 and for the real/log framework used in the identification.","marker":"[6]"},{"why":"Provides A'Campo's original real oriented blowup along simple normal crossings divisors, which rounding generalizes.","marker":"[1]"},{"why":"Foundational paper defining Fontaine–Illusie log structures, from which the log-space definitions are taken.","marker":"[43]"},{"why":"Cited for the proof of Theorem 4.39 on the topology, torsor fibers, and functoriality of rounding.","marker":"[59]"},{"why":"Introduces toroidal embeddings and toroidal varieties, the geometric setting for the boundary-cutting construction.","marker":"[48]"},{"why":"Classical source for the Milnor fibration, whose existence underlies the representative provided by Corollary 4.53.","marker":"[55]"}],"fun_headline_variants":["Rounding log spaces gives canonical singularity links","Real oriented blowups, generalized: rounding for links","Canonical links and Milnor fibers via rounding of log spaces","Polar coordinates to log rounding: canonical links","Ending tubular choices: rounding gives canonical links"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the quoted, unproved results are correct—Nakayama–Ogus's local triviality theorem for roundings, the identification of rounding with real oriented blowup, and Hironaka's resolution theorem—since the canonical links and Milnor fibers are built directly on them.","fun_headline_variants_meta":{"raw":{"variants":["Rounding log spaces gives canonical singularity links","Real oriented blowups, generalized: rounding for links","Canonical links and Milnor fibers via rounding of log spaces","Polar coordinates to log rounding: canonical links","Ending tubular choices: rounding gives canonical links"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000323,"raw_usage":{"total_tokens":1895,"prompt_tokens":1108,"completion_tokens":787,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":724,"completion_tokens_details":{"reasoning_tokens":713}},"tokens_in":724,"tokens_out":787,"duration_ms":8281,"temperature":1.0,"reasoning_tokens":713,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:56:04.141656+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a Brieskorn hypersurface such as $x^2 + y^3 + z^5 = 0$ with its standard toroidal resolution: compute the rounding of the restriction to the exceptional divisor $E$ of the divisorial log structure, and compare the resulting space, with its projection to $E$, to the classical link (a Brieskorn sphere). If the two are not homeomorphic as spaces over $E$, the claim that rounding gives a canonical representative of the link is false.","supporting_citations":[{"cited_title":"Nakayama, A","cited_arxiv_id":null,"evidence_quote":"Supplies the Nakayama–Ogus local triviality theorem (Theorem 4.49), from which the Milnor-fibration representative follows."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the rounding functor and its basic properties; the survey's central object."},{"cited_title":"Kajiwara, C","cited_arxiv_id":null,"evidence_quote":"Quoted for Proposition 4.43, identifying roundings of toric varieties with their real oriented blowups."},{"cited_title":"Arg¨ uz,Real loci in (log) Calabi-Yau manifolds via Kato-Nakayama spaces of toric degenerations","cited_arxiv_id":null,"evidence_quote":"Also quoted for Proposition 4.43 and for the real/log framework used in the identification."},{"cited_title":"Kato, Logarithmic structures of Fontaine-Illusie","cited_arxiv_id":null,"evidence_quote":"Foundational paper defining Fontaine–Illusie log structures, from which the log-space definitions are taken."},{"cited_title":"Ogus, Lectures on logarithmic algebraic geometry","cited_arxiv_id":null,"evidence_quote":"Cited for the proof of Theorem 4.39 on the topology, torsor fibers, and functoriality of rounding."},{"cited_title":"Kempf, F","cited_arxiv_id":null,"evidence_quote":"Introduces toroidal embeddings and toroidal varieties, the geometric setting for the boundary-cutting construction."},{"cited_title":"Milnor, Singular points of complex hypersurfaces","cited_arxiv_id":null,"evidence_quote":"Classical source for the Milnor fibration, whose existence underlies the representative provided by Corollary 4.53."}],"review_version":1}