REVIEW 4 major objections 5 minor
Symmetric lexicographic symmetric-subset reverse search for the enumeration of circuits, cocircuits, and triangulations up to symmetry
T0 review · 4 major / 5 minor · reviewed 2026-07-08 · grok-4.5
Pith's one-line read A reverse-search framework with new lex-minimality tests and pruning rules enumerates cocircuits, circuits, and triangulations up to symmetry, giving the first complete counts for several open cases.
desk verdict Working reverse-search with new lex tests and pruning that delivers first complete orbit counts for several hard cube and triangulation instances, plus a Santos correction. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The symmetric lexicographic symmetric-subset reverse search framework: reverse search over feasible subsets ordered by a lexicographic representative of each orbit under the symmetry group. Lex-minimality of a candidate is decided by the critical-element method or the modified switch-table method; the search tree is reduced by rank-pruning for cocircuits and lex-pruning for triangulations.
What would settle it
An independent enumeration of the cocircuits of the 9-cube or the circuits of the 8-cube by any method that does not rely on the same rank- or lex-pruning, if it produces a different total, would falsify completeness of the pruned search tree.
Extended reading notes
Core claim
The symmetric lexicographic symmetric-subset reverse search framework, using critical-element and modified switch-table tests for orbit lex-minimality plus rank-pruning for cocircuits and lex-pruning for triangulations, correctly enumerates the orbits of cocircuits, circuits, and triangulations and yields the first complete counts for cocircuits of the 9-cube, circuits of the 8-cube, triangulations of Δ5×Δ3, and triangulations of Santos’s 6D 17-point configuration; it further shows that Santos’s 5D 26-point triangulation can be flipped to a regular one in the original configuration.
Load-bearing premise
The application-dependent pruning rules never discard a reverse-search branch that would have produced a lexicographically minimal representative of a new orbit.
Editorial extensions
If this is right
- The number of cocircuits of the 9-cube is known exactly for the first time.
- The number of circuits of the 8-cube is known exactly for the first time.
- All triangulations of the product of a 5-simplex and a 3-simplex can be counted completely.
- All triangulations of Santos’s 6-dimensional 17-point configuration are counted, and its flip-graph component is confirmed to contain only non-regular triangulations.
- Santos’s triangulation of the original 5-dimensional 26-point configuration admits a flip path to a regular triangulation.
Reading between the lines
- Independent recomputation of the new orbit counts by a non-pruned or differently pruned method would provide a strong external check on the completeness claim.
- The sensitivity of the 26-point example to mild coordinate change suggests that claims of disconnected flip graphs should be tested under generic perturbations of the point set.
- The critical-element and switch-table lex-minimality tests may transfer directly to orbit enumeration of other combinatorial objects that admit reverse-search formulations, such as matroid bases.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript develops a symmetric lexicographic reverse-search framework for enumerating orbits of feasible subsets of a finite ground set, specialized to cocircuits, circuits, and triangulations of point configurations. It contributes two lex-minimality tests (critical-element and modified switch-table), two application-dependent pruning rules (rank-pruning for cocircuits; lex-pruning for triangulations), and a C++ realization inside TOPCOM. With these tools the authors recompute known benchmarks substantially faster and report first complete orbit counts for cocircuits of the 9-cube, circuits of the 8-cube, all triangulations of Δ5×Δ3, and all triangulations of Santos’s 6-dimensional 17-point configuration; they further claim that the flip-graph component of the latter is purely non-regular and that Santos’s 5-dimensional 26-point triangulation can be heuristically flipped to a regular triangulation in the original configuration (but not in a mild modification).
Significance. If the completeness arguments for the pruning rules and the lex-minimality tests hold, the work supplies both a reusable algorithmic framework and several long-standing missing exact counts in computational discrete geometry. The open implementation in TOPCOM, the large speed-ups on classical benchmarks, and the corrected picture of Santos’s 5D example are concrete, reusable contributions. The new orbit numbers for the 9-cube, 8-cube, Δ5×Δ3 and the 6D Santos configuration would become standard reference data; the purely non-regular component claim is of independent interest for the theory of secondary polytopes and flip-graph connectivity.
major comments (4)
- [Framework analysis / pruning sections (rank-pruning, lex-pruning)] The headline exact counts (9-cube cocircuits, 8-cube circuits, Δ5×Δ3 triangulations, Santos 6D 17-point triangulations) rest on the claim that rank-pruning and lex-pruning never discard a reverse-search ancestor of a lexicographically minimal orbit representative. Reverse-search correctness requires that the parent map induce a spanning forest on the orbit-representative graph; any pruning predicate must therefore be ancestor-closed with respect to every lex-minimal feasible set. The abstract asserts that these rules are “analyzed,” yet the argument is necessarily application-specific. A self-contained proof (or a clearly isolated lemma) that whenever S is a lex-minimal cocircuit/circuit/triangulation every reverse-search ancestor of S also survives the respective pruning predicate is load-bearing and must be checkable without reconstructing the whole search tree. If that argument is inc
- [Santos 6D 17-point results] The same completeness underpins the claim that Santos’s 6D component is purely non-regular: exhaustive component enumeration plus regularity checks is only exhaustive if no lex-minimal triangulation was pruned away. An independent sanity check—e.g., comparison against an unpruned run on a smaller isomorphic instance, or an explicit certificate that every regular triangulation found by other means appears in the enumerated set—would substantially strengthen the claim.
- [Santos 5D 26-point discussion] For the 5D 26-point configuration the manuscript reports that Santos’s triangulation “can be heuristically flipped to a regular triangulation.” A heuristic flip sequence is not a proof of connectivity. Either an explicit, machine-verifiable sequence of flips (with intermediate triangulations and regularity certificates) should be supplied, or the claim should be weakened to “a heuristic search produced a path.” The distinction matters because the paper presents the observation as detection of a flaw in the literature.
- [Lex-minimality tests (critical-element, modified switch-table)] The critical-element and modified switch-table lex-minimality tests are also load-bearing for orbit correctness. Their soundness proofs should be stated with the precise group action, the definition of the critical element / switch table, and the argument that a non-minimal set is always rejected while every minimal set is accepted. Cross-reference to the classical switch-table literature and a clear statement of what is new versus what is inherited would help the reader assess novelty and correctness risk.
minor comments (5)
- [Title] The title is dense and nearly tautological (“Symmetric lexicographic symmetric-subset reverse search”). A shorter working title that foregrounds the three applications would improve discoverability.
- [Computational results] Benchmark tables should report wall-clock time, number of nodes expanded, and the precise TOPCOM / compiler / hardware configuration so that the “large margin” speed-ups can be reproduced.
- [Preliminaries / notation] Notation for the group action, orbit representatives, and the parent map of reverse search should be fixed once in a preliminary section and reused uniformly; occasional shifts between “symmetric subset,” “orbit,” and “lex-minimal representative” slow reading.
- [Abstract / results summary] The abstract’s phrase “among others” for new numbers is vague; a short table of all newly obtained exact counts (with OEIS or literature pointers where applicable) would be more useful.
- [Software / reproducibility] Clarify whether the C++ code and the exact input files for the new enumerations will be deposited with the paper or only inside the TOPCOM distribution; permanent archival of the Santos instances is especially important.
Simulated Author's Rebuttal
We thank the referee for a careful and constructive report. The four major comments correctly identify the load-bearing correctness claims of the paper (ancestor-closed pruning, soundness of the two lex-minimality tests, and the two Santos applications). We agree that these arguments must be isolated, self-contained, and checkable without reconstructing the search tree, and that the Santos 5D claim must be stated with a precision matching the evidence. In the revision we will supply the missing lemmas, expand the soundness proofs, add independent sanity checks for the 6D enumeration, and either certify or carefully weaken the 5D connectivity statement. We believe these changes fully address the recommendation of major revision while preserving the algorithmic contributions and the new exact counts.
read point-by-point responses
-
Referee: The headline exact counts rest on the claim that rank-pruning and lex-pruning never discard a reverse-search ancestor of a lex-minimal orbit representative. A self-contained proof (or clearly isolated lemma) that whenever S is a lex-minimal cocircuit/circuit/triangulation every reverse-search ancestor of S also survives the respective pruning predicate is load-bearing and must be checkable without reconstructing the whole search tree.
Authors: We agree that ancestor-closure of the pruning predicates is load-bearing for reverse-search correctness and that the present exposition does not isolate the argument sufficiently. In the revision we will add two explicit lemmas (one for rank-pruning of cocircuits, one for lex-pruning of triangulations) stating: if S is a lexicographically minimal feasible set in its orbit and T is any reverse-search ancestor of S under the parent map of the framework, then T satisfies the pruning predicate. Each lemma will be proved from the definitions of the parent map, the lex order, and the application-specific predicate alone, without reference to the global search tree. The surrounding text will make clear that these lemmas, together with the already-stated spanning-forest property of the parent map on orbit representatives, imply that no lex-minimal representative is lost. We will also add a short remark explaining why the same argument does not automatically transfer to circuits (where we do not claim an analogous pruning rule). revision: yes
-
Referee: The same completeness underpins the claim that Santos’s 6D component is purely non-regular. An independent sanity check—e.g., comparison against an unpruned run on a smaller isomorphic instance, or an explicit certificate that every regular triangulation found by other means appears in the enumerated set—would substantially strengthen the claim.
Authors: We accept the request for independent corroboration. In the revision we will (i) report an unpruned reverse-search run on a symmetry-reduced isomorphic copy of a smaller face of the configuration (or on a known subconfiguration whose regular triangulations are independently enumerable), confirming that the pruned and unpruned orbit counts coincide; (ii) verify that every regular triangulation of the 17-point set that can be produced by TOPCOM’s existing regular-triangulation engines appears among the enumerated orbits; and (iii) record the regularity certificates (height vectors or secondary-cone witnesses) for all regular orbits found. These checks do not replace the ancestor-closure lemmas of the previous point, but they give an independent consistency test of the claim that the enumerated component contains no regular triangulation. revision: yes
-
Referee: For the 5D 26-point configuration the manuscript reports that Santos’s triangulation “can be heuristically flipped to a regular triangulation.” A heuristic flip sequence is not a proof of connectivity. Either an explicit, machine-verifiable sequence of flips (with intermediate triangulations and regularity certificates) should be supplied, or the claim should be weakened to “a heuristic search produced a path.” The distinction matters because the paper presents the observation as detection of a flaw in the literature.
Authors: The referee is right: a heuristic path is not a connectivity proof, and the present wording overstates the logical status of the observation. We will revise the abstract and the corresponding section as follows. If, by the time of revision, we can extract and verify a complete flip sequence (each intermediate triangulation written out, each flip certified by a bistellar move, and the terminal triangulation certified regular by a height vector), we will supply that sequence as supplementary data and keep a carefully worded claim of connectivity in the original configuration. Otherwise we will weaken the language throughout to: “a heuristic flip search produced a path from Santos’s triangulation to a regular triangulation in the original point configuration; the same heuristics failed on a mild modification.” We will also remove or soften the phrase “a flaw has been detected,” replacing it by a statement that the computational evidence suggests the original configuration does not separate Santos’s triangulation from the regular component, contrary to what one might have expected from the literature. In either case the distinction between certified path and heuristic evidence will be made explicit. revision: yes
-
Referee: The critical-element and modified switch-table lex-minimality tests are also load-bearing for orbit correctness. Their soundness proofs should be stated with the precise group action, the definition of the critical element / switch table, and the argument that a non-minimal set is always rejected while every minimal set is accepted. Cross-reference to the classical switch-table literature and a clear statement of what is new versus what is inherited would help the reader assess novelty and correctness risk.
Authors: We agree that the soundness arguments must be self-contained and that the novelty boundary should be drawn more sharply. In the revision we will reorganize the relevant section as follows. (1) We fix the group action (the natural action of the symmetry group G of the ground set on its power set) and the total order used for lexicographic comparison. (2) For the critical-element test we give a precise definition of the critical element of a set S, prove that if S is not lex-minimal then the test rejects S, and prove that every lex-minimal S is accepted. (3) For the modified switch-table test we recall the classical switch-table construction (with explicit citations), state exactly which modifications we introduce, and prove soundness of the modified test under the same group action. (4) A short comparison paragraph will list which ingredients are inherited from the classical literature and which (the critical-element criterion and the specific modifications of the switch table needed for reverse-search compatibility) are new. These changes make both correctness and novelty independently checkable. revision: yes
Circularity Check
No significant circularity: reverse-search enumeration of externally defined combinatorial objects; new orbit counts are computational outputs, not definitional or fitted restatements.
full rationale
The paper develops and analyzes a symmetric lexicographic reverse-search framework (critical-element and modified switch-table lex-minimality tests; rank-pruning for cocircuits; lex-pruning for triangulations) and applies it in TOPCOM to enumerate orbits of cocircuits, circuits, and triangulations of given point configurations. The objects counted—circuits, cocircuits, triangulations—are standard, externally defined combinatorial structures; the reported numbers (9-cube cocircuits, 8-cube circuits, triangulations of Δ5×Δ3, Santos’s 6D 17-point configuration, etc.) are outputs of an exhaustive search algorithm, not quantities forced by a fitted parameter or by a definition that already encodes the answer. Known benchmarks are recomputed faster and used as external checks; new counts are first-time enumerations, not renamings of prior closed forms. Self-use of TOPCOM is implementation infrastructure for running the search, not a load-bearing self-citation that substitutes for the correctness argument. Completeness under application-dependent pruning is a correctness/soundness claim about the search tree (whether pruned ancestors can still reach every lex-minimal orbit representative), not a circular reduction of the counts to their own inputs. No uniqueness theorem is imported from prior author work to forbid alternatives; no ansatz is smuggled in via self-citation; no free constants are fitted and then re-presented as predictions. The derivation chain is therefore self-contained algorithmic analysis plus computation against external combinatorial definitions.
Assumptions & free parameters
assumptions (5)
- domain assumption Avis–Fukuda reverse search correctly enumerates a finite set when a valid local search and adjacency oracle are given.
- standard math A finite group acts on the ground set and induces well-defined orbits on feasible subsets; lexicographic order selects a unique representative per orbit.
- domain assumption Circuits and cocircuits are those of the oriented matroid (or cube / point configuration) under the usual chirotope / signed-circuit axioms.
- domain assumption A triangulation is a maximal collection of full-dimensional simplices whose union is the convex hull and whose intersections are common faces.
- ad hoc to paper Rank-pruning (cocircuits) and lex-pruning (triangulations) never eliminate a branch that would yield a new lex-minimal orbit representative.
Cite this review
Pith. "Pith review of Symmetric lexicographic symmetric-subset reverse search for the enumeration of circuits, cocircuits, and triangulations up to symmetry." pith.science (2026). https://pith.science/paper/QBRQPTJZ
@misc{pith2026260705967,
author = {Pith},
title = {Pith review of: Symmetric lexicographic symmetric-subset reverse search for the enumeration of circuits, cocircuits, and triangulations up to symmetry},
year = {2026},
howpublished = {\url{https://pith.science/paper/QBRQPTJZ}},
note = {Machine review of arXiv:2607.05967}
}
read the original abstract
This paper introduces, analyzes, and applies variants of the enumeration framework symmetric lexicographic symmetric-subset reverse search for the enumeration of symmetric feasible subsets of a finite set up to symmetry. The framework is implemented in detail for three applications: cocircuits, circuits, and triangulations of point configurations. There are two new methods presented and analyzed to check the lexicographic minimality of a subset in its orbit: the critical-element method and the modified switch-table method. Moreover, new application-dependent methods to reduce the number of necessary enumeration nodes are introduced: rank-pruning for cocircuits and lex-pruning for triangulations. With a C++-implementation of the ideas in the software package TOPCOM, in all three applications known benchmarks can be computed faster by a large margin. The following new numbers could be computed for the first time (among others): the number of cocircuits of the 9-cube, the number of circuits of the 8-cube, and the number of all triangulations of the product of a 5- and a 3-simplex, as well as the number of all triangulations of a point configuration in dimension six with 17~points with disconnected flip-graph (constructed by Santos). Moreover, for Santos's triangulation it has computationally been checked that its flip-graph component is indeed purely non-regular. Furthermore, in another instance in dimension five with 26 points (also constructed by Santos), a flaw has been detected: Santos's triangulation can be heuristically flipped to a regular triangulation in the original point configuration. In a mildly modified version of the point configuration, the heuristics cannot flip Santos's triangulation to a regular triangulation anymore.
Reviewed July 8, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.