{"id":"4035abbc-6c8d-4243-9089-4c5bc1773434","arxiv_id":"1909.02613","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"All linearly damped Newton equations of a certain broad class are shown to fit a contact-geometry framework, which yields new higher-order geometric integrators with derived error estimates, tested on Kepler, spin-orbit, and Lane-Emden problems.","lead":"This paper rewrites a broad family of damped motion equations, including satellite drag, spin-orbit rotation, and stellar structure models, into a geometric form that supports better numerical simulation. The resulting integrators stay stable at large time steps where standard Runge-Kutta methods fail, and they come with formulas for estimating the numerical error.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The spin-orbit contact Hamiltonian printed in Eq. (35) does not generate the stated Eq. (34); the potential sign is inconsistent with the paper's own Proposition 1.","rationale":"The central Proposition 3 is a direct and correct verification, and the splitting construction in Section 3.2 is sound; the order conjecture in Remark 1 is an acknowledged gap that mainly affects the 'Variational 4th' comparison in Section 4.1.2. The more concrete problem appears in Section 4.2: the printed contact Hamiltonian for the spin-orbit model is inconsistent with the stated Newton equation. Substituting Eq. (35) into Proposition 1 gives θ̈ = −∂(N_z/C)/∂θ − (Ċ/C)θ̇, while Eq. (34) requires θ̈ = N_z/C − (Ċ/C)θ̇; these match only if ∂N_z/∂θ = N_z, which the triaxial torque does not satisfy. The later ∂V/∂q expression in Section 4.2 has the correct sign, so the likely explanation is a typo in Eq. (35); nevertheless, the discrepancy must be checked against the archived code before the spin-orbit results are fully accepted. Because the theoretical framework and the other examples stand, and because the fix is localized, this does not warrant changing the CONDITIONAL verdict, but it should be made an explicit condition of acceptance.","tokens_in":21258,"tokens_out":14282,"duration_ms":139069,"concrete_test":"Inspect the archived Zenodo notebook for the spin-orbit example and re-run the Ω = 1, λ = 0.1 case of Figure 9 with two Hamiltonians: H = p²/2 + N_z/C + (Ċ/C)s as printed in Eq. (35), and H = p²/2 + V + (Ċ/C)s with ∂V/∂θ = −N_z/C as used in the Section 4.2 formulas. Comparing the resulting Poincaré sections with each other and with [21, Figure 1] settles whether Eq. (35) is a typo or whether the reported spin-orbit results were produced by a different model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Using Proposition 1, the Hamiltonian H = p^2/2 + N_z(θ,t)/C + (Ċ/C)s in Eq. (35) yields θ̈ = −∂(N_z/C)/∂θ − (Ċ/C)θ̇. Equation (34), C θ̈ + Ċ θ̇ = N_z, is equivalently θ̈ = N_z/C − (Ċ/C)θ̇. Matching these two expressions requires ∂(N_z/C)/∂θ = −N_z/C, which is not an identity for the triaxial torque N_z ∝ Σ W(m/2,e) sin(2θ − mt). The correct contact Hamiltonian should therefore use a potential V(θ,t) with ∂V/∂θ = −N_z/C, which is precisely the form used later in the ∂V/∂q formula of Section 4.2. Consequently either Eq. (35) is a typo and the reported spin-orbit numerics were produced with the corrected potential, or the spin-orbit experiments integrate a different system than Eq. (34). This is a concrete correctness issue in one of the three headline applications, distinct from the acknowledged order conjecture in Remark 1.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Newton equations of the form q̈ + ∂V(q,t)/∂q + f(t)q̇ = 0 and shows that every such system admits a contact Hamiltonian formulation: Proposition 3 verifies directly that H = Σ p_a²/2 + V(q,t) + f(t)s reproduces Eq. (1) through the contact Hamiltonian equations. The authors then construct two families of geometric integrators: higher-order Galerkin contact variational integrators and even-order contact splitting integrators obtained by Yoshida composition. They derive a modified-Hamiltonian backward error analysis and use it to produce explicit, parameter-free local error estimates, which they compare numerically with the observed error. The methods are applied to three celestial-mechanics-type problems: the modified Kepler problem with time-dependent linear drag, the spin-orbit model with time-varying moment of inertia, and the Lane–Emden equation. The numerical sections report that the contact integrators remain stable at time steps where fixed-step RK4 diverges, and that the modified-Hamiltonian estimates track the actual error on the damped oscillator, modified Kepler, and Lane–Emden tests.","tokens_in":21338,"tokens_out":9403,"duration_ms":97695,"significance":"If the identified issues are resolved, this is a useful contribution to geometric numerical integration. The Hamiltonisation of the broad class (1) is clean and self-contained, with Proposition 3 verified directly from the definition of contact Hamiltonian systems. The splitting integrators are standard Strang/Yoshida compositions, and the modified-Hamiltonian error estimates are genuine predictions rather than fitted curves: they match the reported numerics without adjustable parameters. The paper also ships reproducible code (DOI 10.5281/zenodo.3557652), which strengthens the numerical claims. The contact-geometric viewpoint offers a practical unification for damped Newtonian systems in celestial mechanics, and the comparison against RK4 addresses a question of real interest to practitioners. However, two load-bearing points require attention before the results can be accepted as stated: the printed spin-orbit Hamiltonian in Eq. (35) is inconsistent with the equation it claims to represent, and the fourth-order variational integrator used in Section 4.1.2 is labeled fourth order on the strength of a conjecture that the paper itself declares to be unproved.","major_comments":[{"comment":"The contact Hamiltonian in Eq. (35) does not generate the stated equation of motion (34). By Proposition 1, for H = p²/2 + N_z(θ,t)/C + (Ċ/C)s one obtains θ̈ = −∂(N_z/C)/∂θ − (Ċ/C)θ̇, while Eq. (34) is equivalent to θ̈ = N_z/C − (Ċ/C)θ̇. Matching the two expressions requires ∂(N_z/C)/∂θ = −N_z/C, which is not an identity for the triaxial torque N_z ∝ Σ W(m/2,e) sin(2θ − mt). The correct Hamiltonian must use a potential V(θ,t) satisfying ∂V/∂θ = −N_z/C, and this is precisely the form used later in §4.2 through the expression for ∂V/∂q. Please correct Eq. (35) and state explicitly which system the spin-orbit experiments integrate.","section":"§4.2, Eq. (35)"},{"comment":"The order formula min(2ℓ,u) for the Galerkin contact variational integrators is conjectured, with only numerical evidence from the symplectic analogue [30] and a proof declared future work. Nevertheless, §4.1.2 and Figures 6–8 use a 'Variational 4th' integrator as a fourth-order method in the headline stability and accuracy comparisons. If the actual order is lower, the conclusions drawn from those comparisons would need revision. The paper should either provide a proof of the order statement in the contact case, or explicitly qualify the label and soften the claims that rely on it.","section":"§3.1, Remark 1; §4.1.2"}],"minor_comments":[{"comment":"The captions use γ for the damping coefficient that Eq. (27) and the surrounding text call α; please unify the notation.","section":"Figures 1–2"},{"comment":"The decimal formatting of the coefficients is inconsistent (varying spaces and digit counts), which makes reproduction from the table unnecessarily error-prone; please format all entries uniformly.","section":"Table 1"},{"comment":"In the explicit formula for ∆H′, the term p ∂V/∂q and similar products are written without summation indices; if the multi-dimensional case is intended, please add the sum over a, or state that the formula is for the scalar case.","section":"§3.2.3, Eq. (21)"},{"comment":"A direct comparison with [21, Figure 5] is described, but the paper does not report the parameter values used in [21] or the precise values used for the comparison aside from C̃=1, ν=1, B−A and e; please add these values so the reproducibility of Figure 10 does not depend on the companion code alone.","section":"§4.2.1"}],"recommendation":"major_revision","confidential_remarks":"The spin-orbit issue in Eq. (35) is, in my reading, an error in the displayed Hamiltonian rather than a defect in the numerical implementation, since the later ∂V/∂q expression matches the correct potential. Still, a published paper cannot present Eq. (35) as the Hamiltonian of Eq. (34) without correction. The order conjecture for the variational integrator is an acknowledged limitation, but it is load-bearing for the 'fourth order' label used in the numerical comparisons; I would ask the authors to either prove it or clearly restrict the claim. I saw no citation-pattern concerns: the self-citations [5] and [38] are appropriate as the sources of the contact-Hamiltonian and discrete-Herglotz frameworks."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper deserves a real referee. It does something new and useful—contact Hamiltonisation of a broad class of damped Newtonian systems, plus explicit integrators with derived error estimates—and it is mostly honest about what is proven and what is not.\n\nThe new material is in Section 3: the higher-order Galerkin contact variational integrators and the splitting contact integrators with the explicit modified Hamiltonian ΔH′ for time-dependent systems. Proposition 3 is a direct check, and the error estimates (31), (38)–(40) are parameter-free predictions that match the reported numerics on the damped oscillator and modified Kepler problem without fitted constants. That is real evidence. The Lane–Emden section is also refreshingly candid: the authors flag that Proposition 10 breaks down due to the singularity, and still reproduce the exact n=0,1,5 solutions without special treatment. The Zenodo archive with notebooks helps reproducibility.\n\nSoft spots, in proportion:\n\n- The order of the 'Variational 4th' integrator rests on the conjecture in Remark 1, not a proof. The paper leans on this integrator in the Kepler stability comparisons. They flag it, but it is a real gap; the comparisons involving that integrator should be read accordingly.\n\n- The spin-orbit section is qualitatively compared to [21] with unknown reference parameters. They admit this. It weakens the claim of 'great accuracy' but does not undo the qualitative message.\n\n- There is a concrete sign error in the printed Hamiltonian for the spin-orbit model. Equation (35) gives H = p²/2 + N_z/C + (Ċ/C)s, but Proposition 1 applied to this H yields θ¨ = −∂(N_z/C)/∂θ − (Ċ/C)θ̇, whereas equation (34) requires θ¨ = N_z/C − (Ċ/C)θ̇. Matching requires ∂V/∂θ = −N_z/C, not V = N_z/C. The later ∂V/∂q formula in Section 4.2 has the correct sign, so this is almost certainly a typo, but the authors need to confirm the reported spin-orbit numerics were produced with the corrected Hamiltonian. A typo in a headline application is not fatal, but it is exactly the kind of thing that needs fixing before publication.\n\n- Baselines are thin: fixed-step RK4 for Kepler, no direct quantitative comparison to the symplectic algorithm in [21]. Minor.\n\nThe citation pattern is fine; using their own earlier work [5, 38] to build the framework is legitimate.\n\nWho this is for: anyone working on geometric integrators for dissipative systems or on the contact formulation of classical mechanics. It deserves peer review—with the sign issue resolved and the order conjecture either proven or explicitly labeled as unsupported for the 4th-order results, I would be comfortable seeing it in print. Send it to referees.","headline":"Solid, useful paper on contact integrators; the spin-orbit Hamiltonian as printed has a sign error, and the 4th-order variational integrator rests on an unproved conjecture.","tokens_in":22013,"tokens_out":5613,"would_cite":true,"duration_ms":49239,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65D30","34K28","34A26"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that every Newton equation with velocity-linear time-dependent damping—including the modified Kepler problem, the spin-orbit model, and the Lane-Emden equation—is the flow of a contact Hamiltonian, and builds…","keywords":["contact geometry","geometric integrators","modified Kepler problem","spin-orbit model","Lane-Emden equation","contact Hamiltonian systems","Herglotz variational principle","splitting methods"],"falsifier":"Integrate a contact system with a known exact solution (e.g., the damped harmonic oscillator $H=p^2/2+q^2/2+\\alpha s$) with the 'Variational 4th' integrator at several step sizes $\\tau$ and fit the observed error slope: if the order is not 4, the unproved order conjecture fails and the Section 4.1.2 comparisons involving that integrator would need revision; separately, rerun the modified-Kepler large-$\\tau$ experiment with fixed-step RK4 at a much smaller step until convergence to see whether the claimed stability advantage is about step-size regime rather than structure preservation.","tokens_in":20895,"feed_emoji":"🛰️","tokens_out":7486,"duration_ms":68981,"temperature":0.7,"pith_summary":"This paper claims that the large class of Newton equations with time-dependent damping linear in velocity—covering the modified Kepler problem, the spin-orbit model, and the Lane-Emden equation—can be written as contact Hamiltonian systems, so the geometric toolkit of Hamiltonian mechanics extends to them. It introduces the contact Hamiltonian $H=\\sum_a p_a^2/2+V(q,t)+f(t)s$, whose flow restricts to the given Newton equation on the physical variables while the extra variable $s$ encodes the damping. On this basis the paper constructs two families of integrators that preserve the contact structure: splitting-based Hamiltonian integrators of any even order, and higher-order variational integrators from a discrete Herglotz principle. Numerical experiments show these integrators staying stable at step sizes where fixed-step Runge-Kutta diverges, reproducing known spin-orbit phase portraits and resonance capture, and handling the singular Lane-Emden equation without special treatment.","feed_headline":"Damped celestial orbits are contact Hamiltonian flows","feed_subtitle":"A single contact Hamiltonian unifies the modified Kepler, spin-orbit, and Lane-Emden equations and yields stable integrators.","key_machinery":"The central object is the contact Hamiltonian $H(p,q,s,t)=\\sum_a p_a^2/2+V(q,t)+f(t)s$ on an exact contact manifold with contact form $\\eta=ds-\\sum_a p_a dq^a$; the extra coordinate $s$ absorbs the damping term $f(t)\\dot q$ because Hamilton's equations become $\\dot q^a=p_a$, $\\dot p_a=-\\partial V/\\partial q^a-f(t)p_a$, $\\dot s=\\sum_a p_a^2/2-V-f(t)s$, and the first two reproduce the target Newton equation. The integrators are built by splitting this Hamiltonian into exactly integrable pieces $A=f(t)s$, $B=V(q,t)$, $C=\\sum_a p_a^2/2$, and composing their flows symmetrically; a Baker-Campbell-Hausdorff calculation yields the modified Hamiltonian $\\tilde H'=H+\\tau^2\\Delta H'+O(\\tau^4)$ that supplies local error estimates. On the Lagrangian side, the machinery is the Galerkin discretisation of the discrete Herglotz variational principle, using polynomials of degree $\\ell$ and a Runge-Kutta method of order $u$, producing variational integrators conjectured to have order $\\min(2\\ell,u)$.","core_discovery":"The paper establishes that every equation $\\ddot q+\\partial V(q,t)/\\partial q+f(t)\\dot q=0$ is the flow of the contact Hamiltonian $H=\\sum_a p_a^2/2+V(q,t)+f(t)s$, giving a uniform Hamiltonisation for the modified Kepler problem, the spin-orbit model, and the Lane-Emden equation without the reparameterisation required by earlier symplectic Hamiltonisation attempts. It then builds two families of structure-preserving integrators—splitting-based contact Hamiltonian integrators of arbitrary even order with explicit modified-Hamiltonian error bounds, and higher-order variational integrators from a Galerkin discretisation of Herglotz' principle—and shows numerically that these remain stable at large time steps where fixed-step RK4 diverges, reproduce spin-orbit Poincaré sections and resonance capture, and integrate the singular Lane-Emden equation from its natural initial conditions.","pith_inferences":["If the order conjecture is proved, the Galerkin-Herglotz construction would give a uniform recipe for arbitrary-order variational integrators for any contact Hamiltonian, including non-separable ones such as the damped Schwarzschild geodesics the paper mentions, where splitting methods cannot be applied.","The modified-Hamiltonian error estimate suggests a practical adaptive step-size or initial-condition correction: since the leading error term is known explicitly, one could subtract it or choose modified initial data to reduce drift, especially for the Lane-Emden first step where the paper gives explicit initial errors.","The Hamiltonisation is not merely formal: because the flow is contact, one could apply contact analogues of weak-KAM and Aubry-Mather theory to the modified Kepler and spin-orbit models to study capture into resonances and dissipation-driven transitions analytically, not just numerically."],"forward_implications":["Every system in the class (1) gains a Hamiltonian formulation in one extra dimension, so tools from Hamiltonian dynamics—variational principles, invariants, and periodic-orbit theory—become available for damped Newtonian systems.","The second-order contact Hamiltonian integrator with modified Hamiltonian $\\tilde H'$ produces explicit local error estimates that match observed errors in the damped oscillator and perturbed Kepler tests.","Contact Hamiltonian splitting integrators can be built to any even order from exactly integrable pieces; for separable Hamiltonians of the form (6) this yields explicit, structure-preserving methods.","For the modified Kepler problem with time-dependent drag, the contact integrators remain bounded and stable at time steps where fixed-step RK4 diverges, with Hamiltonian integrators showing more precession but greater stability at large $\\tau$ than the variational ones.","For the singular Lane-Emden equation, the contact Hamiltonian integrator handles the natural initial condition without special treatment, and the modified Hamiltonian predicts bounded first-step errors despite the failure of Proposition 10."],"supporting_citations":[{"why":"Supplies the discrete Herglotz variational principle and the second-order contact variational integrator used as a baseline and starting point.","marker":"[38]"},{"why":"Provides the numerical evidence from the symplectic Galerkin construction on which the order conjecture in Remark 1 rests.","marker":"[30]"},{"why":"Supplies the construction of arbitrary even-order integrators from a second-order base via the coefficients used in Propositions 7 and 8.","marker":"[41]"},{"why":"Grounds the splitting method and exact-integrability conditions used to build the contact Hamiltonian integrators.","marker":"[29]"},{"why":"Introduces contact Hamiltonian mechanics and the damped harmonic oscillator Hamiltonian used as a test case.","marker":"[5]"},{"why":"Provides the spin-orbit Hamiltonian formulation whose Poincaré sections and resonance capture the paper reproduces.","marker":"[21]"},{"why":"Supplies the Cayley coefficients and resonance terminology used in the spin-orbit model.","marker":"[10]"},{"why":"Provides the modified Kepler problem with linear drag as a motivating example of the class (1).","marker":"[27]"},{"why":"Supplies exact Lane-Emden solutions for n=0, 1, and 5 used to validate the numerical error estimates.","marker":"[31]"}],"fun_headline_variants":["Contact Hamiltonians unify damped celestial equations","Stable integrators for damped orbits via contact geometry","A single contact flow for Kepler, spin-orbit, and Lane-Emden","Contact geometry yields robust integrators for damped systems"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fourth-order label on the 'Variational 4th' integrator depends on an unproved conjecture—supported only by numerical evidence from the symplectic version—that the Galerkin-Herglotz construction has order $\\min(2\\ell,u)$; if the true order is lower, every accuracy and stability comparison that uses that integrator would need to be re-examined.","fun_headline_variants_meta":{"raw":{"variants":["Contact Hamiltonians unify damped celestial equations","Stable integrators for damped orbits via contact geometry","A single contact flow for Kepler, spin-orbit, and Lane-Emden","Contact geometry yields robust integrators for damped systems"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000165,"raw_usage":{"total_tokens":1178,"prompt_tokens":799,"completion_tokens":379,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":415,"completion_tokens_details":{"reasoning_tokens":310}},"tokens_in":415,"tokens_out":379,"duration_ms":4512,"temperature":1.0,"reasoning_tokens":310,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:47:25.302448+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Integrate a contact system with a known exact solution (e.g., the damped harmonic oscillator $H=p^2/2+q^2/2+\\alpha s$) with the 'Variational 4th' integrator at several step sizes $\\tau$ and fit the observed error slope: if the order is not 4, the unproved order conjecture fails and the Section 4.1.2 comparisons involving that integrator would need revision; separately, rerun the modified-Kepler large-$\\tau$ experiment with fixed-step RK4 at a much smaller step until convergence to see whether the claimed stability advantage is about step-size regime rather than structure preservation.","supporting_citations":[{"cited_title":"& Seri M.Contact variational integrators","cited_arxiv_id":null,"evidence_quote":"Supplies the discrete Herglotz variational principle and the second-order contact variational integrator used as a baseline and starting point."},{"cited_title":"& Saake N","cited_arxiv_id":null,"evidence_quote":"Provides the numerical evidence from the symplectic Galerkin construction on which the order conjecture in Remark 1 rests."},{"cited_title":"Physics letters A, 150:262–268, 1990","cited_arxiv_id":null,"evidence_quote":"Supplies the construction of arbitrary even-order integrators from a second-order base via the coefficients used in Propositions 7 and 8."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Grounds the splitting method and exact-integrability conditions used to build the contact Hamiltonian integrators."},{"cited_title":"& Tapias D.Contact Hamiltonian mechanics","cited_arxiv_id":null,"evidence_quote":"Introduces contact Hamiltonian mechanics and the damped harmonic oscillator Hamiltonian used as a test case."},{"cited_title":"& Celletti A.Hamiltonian formu- lation of the spin-orbit model with time-varying non-conservative forces","cited_arxiv_id":null,"evidence_quote":"Provides the spin-orbit Hamiltonian formulation whose Poincaré sections and resonance capture the paper reproduces."},{"cited_title":"Zeitschrift für angewandte Mathe- matik und Physik ZAMP, 41:174–204, Mar 1990","cited_arxiv_id":null,"evidence_quote":"Supplies the Cayley coefficients and resonance terminology used in the spin-orbit model."},{"cited_title":"& Rebelo C.Dynamics of Kepler problem with linear drag","cited_arxiv_id":null,"evidence_quote":"Provides the modified Kepler problem with linear drag as a motivating example of the class (1)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies exact Lane-Emden solutions for n=0, 1, and 5 used to validate the numerical error estimates."}],"review_version":1}