{"id":"0169af1e-08dd-4ede-a7cf-e6371a906447","arxiv_id":"1908.07152","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A survey classifying four cut types against five polyhedron classes, producing a table of solved, open, and impossible unfolding problems.","lead":"This short survey organizes twenty questions about cutting open polyhedra into flat nets, sorting which are solved, which have counterexamples, and which remain open. It is useful as a map of the field for researchers choosing open problems in computational geometry.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Table 1's 'Orthogonal—Anycut-Unf' ✓ appears to exceed the cited sources: [BDD+98] covers only some classes, [DDFO17] only genus-2, and [DFO07] is an epsilon-approximation, so the central map may mark an open cell as solved.","rationale":"The paper is a survey, so its central contribution is the 4x5 status table. The reader already flagged that the table rests on citations; the stress-test sharpens this into a concrete, checkable mismatch: at least the Orthogonal—Anycut-Unf entry, and the inherited Polycubes entry, appear to be asserted on the basis of cited works whose scopes are narrower than the cell. This is load-bearing because one wrong ✓ invalidates the 'complete map' claim. The taxonomy and exposition are clear, and the concern could be resolved by a citation audit rather than new mathematics, so a conditional accept is appropriate. If the audit reveals a hidden theorem covering all orthogonal polyhedra, the reader's ACCEPT verdict should stand unchanged.","tokens_in":2401,"tokens_out":12891,"duration_ms":127206,"concrete_test":"Run a citation-scope audit on the disputed cell: locate the exact theorem statement in [BDD+98], [DDFO17], and [DFO07] and check its hypotheses. Does any theorem conclude that every orthogonal polyhedron—all genera and arbitrary face complexity—admits an anycut net? If the strongest available theorem is restricted to 'some classes' or 'genus-2' or gives only epsilon-unfolding, change 'Orthogonal—Anycut-Unf' from ✓ to ? and revisit 'Polycubes—Anycut-Unf' accordingly. Repeating this audit for the remaining non-? cells would provide a cell-by-cell verification of Table 1.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central claim is Table 1's complete status map. No proof is given for any ✓/✗ cell, so each entry must be exactly supported by a citation. The cited orthogonal-polyhedra works do not visibly supply the claimed full-class results. [BDD+98] is titled 'Unfolding some classes of orthogonal polyhedra'; [DDFO17] is restricted to genus-2 polyhedra with linear refinement; [DFO07] proves epsilon-unfolding, an approximation result. None of these states 'every orthogonal polyhedron admits an anycut net', which is what the 'Orthogonal—Anycut-Unf' ✓ asserts. Since every polycube is an orthogonal polyhedron, the 'Polycubes—Anycut-Unf' ✓ inherits this gap. The same scope-mismatch risk applies to other cells: for example, the only cited unzipping counterexample [DDEO19] concerns polycubes, so 'Convex—Edge-Unzip' ✗ would need a separate source. If the orthogonal anycut cells are actually open or only known for subclasses, Table 1 no longer accurately summarizes the state of knowledge, and the survey's central objective fails.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper is a short survey that organizes the known status of a 4-by-5 family of questions about unfolding polyhedra: four cut types (edge-unfold, anycut-unfold, edge-unzip, anycut-unzip) versus five polyhedron classes (convex, spherical, nonconvex, orthogonal, polycubes). Its central output is Table 1, which assigns to each of the 20 combinations a status of 'proven true,' 'false (counterexample exists),' or 'open,' reporting that only 6 combinations are unresolved. The paper provides no proofs or proof sketches; it relies entirely on the cited literature.","tokens_in":2597,"tokens_out":9489,"duration_ms":87586,"significance":"If Table 1 is accurate, the paper provides a compact and useful map of a fragmented literature, highlighting which specializations and generalizations of Dürer's problem are settled and which remain open. The categorization of cut types (especially the distinction between 'unfold' and 'unzip') is clear and helpful. However, because the entries are asserted without any per-cell derivation or citation attribution, the paper's value is entirely dependent on the correctness and scope matching of the citations.","major_comments":[{"comment":"The ✓ entries for 'Anycut-Unf' in the Orthogonal and Polycubes rows are not supported by the cited sources. [BDD+98] claims only 'some classes' of orthogonal polyhedra; [DDFO17] treats only genus-2 polyhedra and requires linear refinement; and [DFO07] establishes only an epsilon-version of unfolding. None of these proves that every orthogonal polyhedron (or every polycube) admits a cut along arbitrary curves that yields a single non-overlapping net. Since the table's ✓ is a universal claim, the author must either provide a reference that states the unrestricted theorem or mark the cell as open. This issue is load-bearing because the paper's central claim is that Table 1 accurately summarizes the state of knowledge.","section":"Table 1, row Orthogonal/Polycubes, column Anycut-Unf"},{"comment":"The ✗ entries for edge-unzipping of convex and spherical polyhedra are not backed by the only unzipping counterexample citation in the paper, [DDEO19], which concerns polycubes. An edge-unzipping is a Hamiltonian path in the 1-skeleton, so the claim that not all convex (or spherical) polyhedra can be edge-unzipped requires a separate source constructing a convex polyhedron with no such path. Without that citation or a proof, the negative entries cannot be verified.","section":"Table 1, rows Convex/Spherical, column Edge-Unzip"}],"minor_comments":[{"comment":"The phrase 'any curve on the surface that form a spanning tree of the vertices' is imprecise: the cut set in a general unfolding is a tree on the surface (a cut graph), not literally a spanning tree of the vertex set. Please reword to avoid confusion.","section":"Section 1, definition of anycut-unfold"},{"comment":"The abbreviations 'Anycut-Unf' and 'Anycut-Unzip' should be expanded in the caption or in a footnote to the table for clarity.","section":"Table 1, caption"},{"comment":"The ✗ entry for edge-unfolding of nonconvex polyhedra is plausible, but no specific counterexample is cited; please add a reference (e.g., to a known nonconvex polyhedron with no edge unfolding).","section":"Table 1, row Nonconvex, column Edge-Unf"},{"comment":"There is a typo: 'Prod. Symp. Comput. Geom.' should be 'Proc. Symp. Comput. Geom.'","section":"Reference list, [O'R18]"}],"recommendation":"major_revision","confidential_remarks":"The paper reads as a CCCG abstract rather than a full journal article. If this is intended for a journal, the citation-scope issues in Table 1 must be resolved; otherwise the survey's main deliverable is unverifiable. I would also ask the editor to check whether the 'Orthogonal—Anycut-Unf' result is actually known in the broader literature; if it is open, the total number of open problems is 8 rather than 6, which changes the abstract."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a survey, not a new-theorem paper, and it is honest about that. The 'anycut' terminology is a genuine addition—it gives the field a clean name for the non-edge-cutting direction—and the 4×5 grid does make the literature easier to talk about. If Table 1 is right, it gives the community exactly the kind of compact status map people will steal for talks. The list of six open problems alone is worth having.\n\nThe good news is that the paper is clear and the taxonomy is sensible. The bad news is that the table is the whole product, and it is not auditable from the text. Every ✓ and ✗ is asserted by citation, with no exact theorem statement and no proof sketch. For a map whose entire purpose is to tell the reader what is known, that is a load-bearing weakness, not a cosmetic one.\n\nThe stress-test note about the Orthogonal—Anycut-Unf entry looks right to me. [BDD+98] is 'some classes', [DDFO17] is genus-2, and [DFO07] is epsilon-unfolding. None of those plainly establishes 'every orthogonal polyhedron admits an anycut net'. Same problem shows up in the row for polycubes, which inherits the orthogonal claim, and in Convex—Edge-Unzip, where the only cited counterexample paper is about polycubes, not general convex polyhedra. Maybe the author has other sources in mind, perhaps in [O'R08], but the paper does not say which cited theorem yields which cell. So a reader cannot tell whether the table is a verified inventory or a best-effort guess.\n\nThat said, I would not call this a sloppy paper. The author almost certainly knows the literature far better than I do, and it is possible every entry is correct. But a survey's contract is provenance: if you tell me a problem is closed, you need to point at the exact theorem. As written, the paper asks me to take the whole map on faith.\n\nWho is this for? People entering the area who want a quick orientation, and established researchers who want a compact visual for a talk. It is not a research contribution in the theorem sense, and it should not be judged as one.\n\nMy recommendation for peer review: send it to a referee who knows these sources cold, and ask for a provenance column or endnotes that map each cell to a specific theorem. If that is fixed, the survey earns its place. If not, the table should be flagged as unverified. This deserves referee time, but it needs revision before I would rely on the map.","headline":"A useful survey-style map of unfolding problems, but the central table's solved/open cells are asserted by citation and at least one orthogonal-polyhedra cell appears over-claimed.","tokens_in":3087,"tokens_out":2911,"would_cite":false,"duration_ms":34532,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52B10","05C45","68U05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Unfolding polyhedra: of the 20 combinations of cut type and polyhedron shape, 14 are proven settled and 6 remain open.","keywords":["edge-unfolding","polyhedral nets","unzipping","convex polyhedra","orthogonal polyhedra","polycubes","spanning-tree cuts","computational geometry"],"falsifier":"Locate the original proof behind any single check entry and test whether it covers every polyhedron in the stated class; for example, a convex polyhedron whose every spanning-tree cut yields an overlapping net would overturn the convex anycut-unfold cell, and likewise for any other check.","tokens_in":2197,"feed_emoji":"🧊","tokens_out":11241,"duration_ms":95694,"temperature":0.7,"pith_summary":"This paper surveys the question of when a polyhedron can be cut open and flattened into one non-overlapping piece. It organizes that question into twenty cases: four ways to choose the cuts (edges only, any spanning-tree curve, an edge path, or any curve through all vertices) and five families of polyhedra (convex, spherical, nonconvex, orthogonal, and polycubes). The paper's claim is that fourteen of those twenty cases are currently settled by proof or counterexample, and six are open. Its purpose is to make the boundary of current knowledge explicit, so future work can target exactly the unresolved cells.","feed_headline":"14 of 20 polyhedron-unfolding questions are settled; 6 remain open","feed_subtitle":"A compact map of four cut types against five shape classes shows where proofs, counterexamples, and open frontiers stand.","key_machinery":"The load-bearing object is the status table itself. The four cut types form a ladder: 'edge-unfold' permits only polyhedron edges, 'anycut-unfold' permits any curve forming a vertex-spanning tree, 'edge-unzip' requires the cut edges to form a Hamiltonian path through the vertices, and 'anycut-unzip' relaxes that to any simple curve touching every vertex. The five shape classes range from the restrictive convex and spherical polyhedra to the broad nonconvex class and the right-angled orthogonal and polycube families. By placing each prior result in one cell, the table turns scattered knowledge into a single assertion: exactly these fourteen cells are settled and exactly these six are open.","core_discovery":"The central claim is the correctness of a 4-by-5 table that classifies every combination of cut type and polyhedron class. In the table, a check means every polyhedron in that class can be unfolded with cuts of that type; a cross means at least one polyhedron in the class cannot; a question mark means the case is unresolved. The paper assigns checks to anycut-unfolding for convex, spherical, orthogonal, and polycube polyhedra, and to anycut-unzipping for orthogonal and polycube polyhedra. It assigns crosses to the edge-unzipping row across all five classes and to edge-unfolding for nonconvex and orthogonal polyhedra. The six remaining question marks sit in the edge-unfolding column for convex, spherical, and polycube polyhedra, in the anycut-unfolding cell for nonconvex polyhedra, and in the anycut-unzipping cells for convex and spherical polyhedra.","pith_inferences":["If the table is accurate, the classical edge-unfolding problem is not isolated: its three open cells (convex, spherical, polycubes) all sit in the same column, so a breakthrough for any one shape family may transfer to the others by the same cut method.","The open anycut-unzip cells for convex and spherical polyhedra look like the nearest targets, since the analogous cells for orthogonal and polycube polyhedra already have positive proofs.","A reader could extend the taxonomy by adding a sixth shape class or a fifth cut type; the table's axes are modular and the new open/closed pattern would immediately expose whether the current boundary is an artifact of the chosen categories."],"forward_implications":["Any future proof or counterexample in one of the six open cells will change the table, giving researchers a clear target.","The crosses show that searching for universal edge-unzippings is pointless across all five classes, since counterexamples already exist.","The successful anycut-unfold proofs for orthogonal and polycube families suggest that freeing the cuts from edges is what makes those classes tractable.","The table gives a compact way to test claims: a new result that contradicts a check or a cross must be reconciled with the cited counterexample or proof."],"supporting_citations":[{"why":"Supplies the standard reference for the edge-unfolding problem and the earlier notion of general unfoldings that 'anycut' replaces.","marker":"[DO07]"},{"why":"Formulates the mathematical edge-unfolding question and introduces the unzipping specialization via Hamiltonian paths.","marker":"[She75]"},{"why":"Provides the counterexample that some polycubes have no edge-unzipping, grounding every cross in the edge-unzip row.","marker":"[DDEO19]"},{"why":"Proves anycut-unfolding for level-1 Menger polycubes of arbitrary size, grounding the polycube anycut-unfold check.","marker":"[RALSZ19]"},{"why":"Describes spiral unfoldings of convex polyhedra, grounding the spherical anycut-unfold check.","marker":"[O'R15]"},{"why":"Establishes edge-unfolding of nearly flat convex caps, the recent progress cited for the convex edge-unfold cell.","marker":"[O'R18]"},{"why":"Unfolds some classes of orthogonal polyhedra, grounding the orthogonal anycut-unfold check.","marker":"[BDD+98]"},{"why":"Demonstrates unfoldings of higher-genus orthogonal polyhedra under refinement, supporting the orthogonal anycut-unzip entry.","marker":"[DDFO17]"}],"fun_headline_variants":["Polyhedron unfolding: 14 of 20 cases settled, 6 remain open","Unfolding polyhedra: 4 cut types, 5 classes, 14 solved, 6 open","14 polyhedron-unfolding problems solved, 6 still unsolved","Polyhedron unfolding map: 14 definitive answers, 6 open frontiers","4×5 grid of unfolding cuts: 14 settled, 6 question marks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The table is only as reliable as the cited results it transcribes, and the paper neither proves nor checks the assumptions behind those citations; if one cited theorem applies to a slightly different class of polyhedra or a different cut definition, the corresponding cell is mislabelled.","fun_headline_variants_meta":{"raw":{"variants":["Polyhedron unfolding: 14 of 20 cases settled, 6 remain open","Unfolding polyhedra: 4 cut types, 5 classes, 14 solved, 6 open","14 polyhedron-unfolding problems solved, 6 still unsolved","Polyhedron unfolding map: 14 definitive answers, 6 open frontiers","4×5 grid of unfolding cuts: 14 settled, 6 question marks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000816,"raw_usage":{"total_tokens":3487,"prompt_tokens":768,"completion_tokens":2719,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":384,"completion_tokens_details":{"reasoning_tokens":2610}},"tokens_in":384,"tokens_out":2719,"duration_ms":21098,"temperature":1.0,"reasoning_tokens":2610,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:35:35.539197+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Locate the original proof behind any single check entry and test whether it covers every polyhedron in the stated class; for example, a convex polyhedron whose every spanning-tree cut yields an overlapping net would overturn the convex anycut-unfold cell, and likewise for any other check.","supporting_citations":[],"review_version":1}