version: v0.4 project: name: Sharp continued-fraction bounds from exact digit runs description: >- Proves sharp bounds for both extrema of three consecutive ordinary continued-fraction approximation coefficients. Exact finite runs of every integer central digit at least two have rational four-corner constants, unique coordinate extremizers, strict actual inequalities, and optimality within the prescribed digit classes. The project also proves exact clipping on every closed unit-square rectangle for every real parameter at least one; sharp inverse stability for both extrema for every real parameter at least two, including their equality loci and actual-convergent optimality; fixed-word sharpness at a fixed index; constant-future and exhausted-past boundary results; effective two-sided run-gap estimates; optimal actual constant-digit-window cutoffs; and an executable rational Lean interface. authors: [JD Jones] responsible_maintainers: [JD Jones] license: MIT classification: arxiv: [math.NT] msc2020: ['11J70', '11A55', '26D07'] sources: - title: Sharp continued-fraction bounds from exact digit runs type: original-proof authors: [JD Jones] relationship: other note: >- This project first presents its exact-run classification, sharp inverse estimates, sharp window cutoffs, and accompanying rectangle and quantitative results. Proof.md and Applications.md accompany the same development and are not distinct earlier sources. The original-proof provenance designation does not assert worldwide priority. Classical metallic thresholds and natural-extension geometry are credited below. - title: Sharpening Borel's result in Diophantine approximation authors: [Savita S. Adhin, Ayreena Bakhtawar, Cor Kraaikamp] id: https://arxiv.org/abs/2607.04421v1 type: paper location: Equations (15)-(16), Sections 3-5 relationship: background note: >- Develops the central-digit-one refinement. Section 5 explicitly suggests an arbitrary-digit extension and a refinement of the conjugate maximum bound. That suggestion and the geometric method are prior work; the present package treats exact runs for digits at least two and the quantitative inverse estimate. - title: Generalization of some classical theorems concerning triples of consecutive convergents to simple continued fractions authors: [F. Bagemihl, J. R. McLaughlin] type: paper location: J. Reine Angew. Math. 221 (1966), 146-149 relationship: background note: >- Classical arbitrary-digit minimum threshold, checked through the accessible later literature. Original full text was not retrieved. - title: The conjugate property of the Borel theorem on diophantine approximation authors: [J. Tong] type: paper location: Math. Z. 184 (1983), 151-153 relationship: background note: >- Classical arbitrary-digit maximum threshold, checked as restated in later papers. Original full text was not retrieved. - title: Sharp bounds for symmetric and asymmetric Diophantine approximation authors: [Cor Kraaikamp, Ionica Smeets] id: https://arxiv.org/abs/0806.1457v2 type: paper location: Sections 2-4 relationship: background note: >- Related natural-extension rectangles and boundary-intersection optimization for different approximation quantities. Sources.md gives the broader literature account and access limitations. automation: methods: - method: agent models: [GPT-6] framework: OpenAI Codex tool_setup: OpenAI Codex with Lean and exact arithmetic tools prompting_notes: >- Agents developed and reviewed mathematical arguments, formal proofs, documentation, and verification under JD Jones's direction. - method: other models: [GPT-6 Pro] tool_setup: Mathematical review and independent exact arithmetic prompting_notes: >- Supplied feedback and proposed extensions; Codex agents adapted the exact calculator and formalized the resulting statements. - method: other models: [Gemini-Argon] tool_setup: Preliminary mathematical discussion notes: >- Disclosure.md records automated contributions. Lean tactics construct kernel-checkable proof terms. Exact Python replays are finite checks, not proof oracles or substitutes for unrestricted formal statements. review: status: agent-reviewed notes: >- Separate agent review checked ordinary-convergent semantics, indexing, exact departures, finite-endpoint alternatives, sharpness quantifiers, both stability equality loci, and sharp digit-window quantifiers. Verification.md records the reproducible mechanical checks. No independent human specialist review or worldwide-priority determination is claimed. status: scope: >- The compared contracts use actual ordinary convergents and floor/fract digits of irrational real inputs. Exact-run sharpness allows the finite index to vary. Separate fixed-word sharpness preserves the complete positive finite word and its index. Rectangle optimization covers every real parameter at least one and all closed rectangles in the unit square, including degenerate rectangles and the lower-envelope zero-axis exception. Both inverse inequalities cover every real parameter at least two. Boundary contracts include exhausted finite pasts, the characterization of infinite constant futures, genuine attained metallic-future examples, and exact alternating coefficient formulas. Quantitative contracts give two-sided run-gap bounds and optimal symmetric digit-window thresholds over actual triples with varying finite index. Four-endpoint rectangle cutoffs also force asymmetric actual windows; separate actual optimality for asymmetric windows is not asserted. Rational executable definitions are proved equal to the corresponding real exact-run constants. Verification.md and comparator.json specify the complete compared declarations. sorry_count: 0 sorry_in_definitions: 0 axioms: [propext, Classical.choice, Quot.sound] main_results: - declaration: ContinuedFractionBounds.ordinary_convergent_triple file: ContinuedFractionBounds/ContinuedFractions.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.lower_rectangle_maximum file: ContinuedFractionBounds/RectangleOptimization.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.upper_rectangle_minimum file: ContinuedFractionBounds/UpperRectangleOptimization.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_run_bounds file: ContinuedFractionBounds/ActualRuns.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_run_lower_sharp file: ContinuedFractionBounds/ActualRuns.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_run_upper_sharp file: ContinuedFractionBounds/ActualRuns.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.global_stability file: ContinuedFractionBounds/Stability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.global_stability_eq_iff file: ContinuedFractionBounds/Stability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.stabilityConstant_optimal file: ContinuedFractionBounds/Stability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_stabilityConstant_optimal file: ContinuedFractionBounds/Applications.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.fixed_past_bounds file: ContinuedFractionBounds/FixedPast.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.fixed_past_lower_sharp file: ContinuedFractionBounds/FixedPast.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.fixed_past_upper_sharp file: ContinuedFractionBounds/FixedPast.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.infinite_future_iff_metallic file: ContinuedFractionBounds/MetallicBoundary.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.metallic_future_run_extrema file: ContinuedFractionBounds/MetallicBoundary.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.metallic_theta file: ContinuedFractionBounds/MetallicBoundary.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.run_constants_convergence file: ContinuedFractionBounds/Quantitative.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.small_deficit_digit_window file: ContinuedFractionBounds/DigitWindows.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.global_upper_stability file: ContinuedFractionBounds/UpperStability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.global_upper_stability_eq_iff file: ContinuedFractionBounds/UpperStability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.upperStabilityConstant_optimal file: ContinuedFractionBounds/UpperStability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_upperStabilityConstant_optimal file: ContinuedFractionBounds/ActualUpperStability.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.run_gap_bounds file: ContinuedFractionBounds/RunGapBounds.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.lowerProfile_optimal file: ContinuedFractionBounds/FiberProfiles.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.upperProfile_optimal file: ContinuedFractionBounds/FiberProfiles.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.sharp_window_complement_bounds file: ContinuedFractionBounds/SharpDigitWindows.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.rational_actual_run_bounds file: ContinuedFractionBounds/RationalBounds.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.sharp_window_digit_window file: ContinuedFractionBounds/SharpDigitWindows.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_lower_window_threshold_optimal file: ContinuedFractionBounds/SharpWindowsRealization.lean comparator_config: comparator.json - declaration: ContinuedFractionBounds.actual_upper_window_threshold_optimal file: ContinuedFractionBounds/SharpWindowsRealization.lean comparator_config: comparator.json fidelity: divergences: >- The written account and Lean both number the triple n,n+1,n+2 and the central digit n+2. Closed run cells include rational endpoints to express sharp suprema and infima. Actual infinite futures are irrational, which supplies strictness in terminated future classes. Finite reversed expansions ending in one are retained. The optional empty-past convention is outside the ordinary triples formalized here.