56-Report-Level Methods Appendix Template v1.0 | Chapter 3 Report Structure (Human-Readable + Machine-Readable)

Home ／ Docs-Technical WhitePaper ／ 56-Report-Level Methods Appendix Template v1.0

Chapter 3 Report Structure (Human-Readable + Machine-Readable)

I. Chapter Goals & Scope (Mandatory)

• Fix a one-to-one structure between the human prose and the machine list, so each method entry is reproducible, auditable, and replayable.
• Applies to: the standard presentation and export (YAML/JSON equivalent) of a single method entry in a release report.

II. Human Structure (fixed order, Mandatory)

• Method Name & Positioning: one-sentence definition; in-scope and non-goals.
• Theory Essentials (S/P anchors): key assumptions, boundaries, minimal derivation.
• Mathematical Formulation & Symbol Table: main equations and symbol list; include path gamma(ell) and measure d ell.
• Algorithm & Pseudocode: steps, complexity, numerical stability notes.
• Data & Experimental Design: datasets@version, sampling & control, variable control.
• Metrology & Calibration (M): units/dimensions, calibration flow, error sources & propagation; check_dim=true.
• Evaluation Protocol & Gates: metric definitions, windows & thresholds; gate_<metric><comparator><threshold>@<window>.
• Uncertainty & Tests: CI_95%, power and multiplicity control.
• Ablation & Sensitivity: variable/path/hyperparameter sensitivity and delta_* indicators.
• Implementation Binding (I): API/Schema/function prototypes, compatibility policy.
• Reproducibility & Artifacts: environment, container, script@commit, dataset@version, artifacts & hashes.
• Risks & Rollback Entry: triggers, rollback steps, and restoration gates.
• References & Cross-References: fixed style “See 《 vX.Y》 Ch.x S/P/M/I…”.

III. Machine Structure (fields & minimums, Mandatory)

• method.* (name/scope/anchors)
• math.* (equations/symbols/path & measure/dimensional check)
• algo.* (pseudocode/complexity)
• data.* (datasets/sampling/controls)
• metrology.* (units/dim/calibration/error)
• evaluation.* (metrics/gates/statistics/window)
• uncertainty.* (CI/discount policy)
• ablation.* (plan and highlights)
• impl.* (API/function prototypes/compatibility)
• reproducibility.* (container/scripts/minimal repro command/artifacts)
• risks.* (triggers/rollback/restoration gates)
• references.see[] (Volume + Version + Anchor)

IV. Field & Constraint List (copy-ready)

V. Machine Schema (YAML; JSON-equivalent, Mandatory)

method:

name: ""

scope: { in: "", out: "", non_goals: [] }

anchors: { S: [], P: [], M: [], I: [] }

math:

statements:

- "T_arr = ( ∫ ( n_eff / c_ref ) d ell )"

path: "gamma(ell)"

measure: "d ell"

symbols:

- { name: "n_eff", unit: "1", dim: "1" }

- { name: "c_ref", unit: "m·s^-1", dim: "L T^-1" }

check_dim: true

algo:

pseudocode: |

# steps...

complexity: { time: "O()", space: "O()" }

data:

datasets: ["cmb_set_v3@v3", "lens_v1@v1"]

sampling: "stratified"

controls: ["region", "band"]

metrology:

units: ["s","m","kg"]

calibration: ["Mx-*"]

error_sources: ["instrument","numerical"]

evaluation:

metrics:

- { name: "gate_accuracy", direction: "↑" }

- { name: "gate_latency", direction: "↓" }

gates_hard: ["gate_accuracy>=0.99@7d"]

gates_soft: ["unit_cost<=1.0x@30d"]

stats: { method: "bootstrap", ci: "95%" }

uncertainty:

discount: { k_sigma: 1.0, formula: "s' = clamp(s - k*sigma, 0, 1)" }

ablation:

plan: ["remove_component_X","freeze_param_Y"]

highlights: []

impl:

api: ["/api/v1/methods/{name}/evaluate"]

functions: ["evaluate_method(req: EvalReq) -> EvalReport"]

compatibility: { version_range: "[1.0,2.0)", fallback: "adapter_v1" }

reproducibility:

container: "registry/replay:2025.09@sha256:…"

scripts: ["arrive_time_check.py@a1b2c3"]

repro_cmd: "docker run … replay --suite arrive_time_check.py --dataset cmb_set_v3@v3"

risks:

triggers: ["gate_accuracy<0.98@7d","compat_rate<0.99@replay"]

success_gates: ["gate_accuracy>=0.99@24h","compat_rate>=0.995@replay"]

artifacts: ["yaml","json","pdf"]

references:

see:

- "EFT.WP.Core.Equations v1.1:S20-1"

- "EFT.WP.Core.Metrology v1.0:check_dim"

- "EFT.WP.Core.DataSpec v1.0:I30-2"

VI. Human × Machine Mapping (Mandatory)

VII. Minimal Sample (human abstract × machine snippet, Mandatory)

1. Human abstract:
   • Use the general form: T_arr = ( ∫ ( n_eff / c_ref ) d ell ); path gamma(ell) is the observation segment and d ell the arc-length element; check_dim=true.
   • Hard gate: gate_accuracy>=0.99@7d; soft gate: unit_cost<=1.0x@30d.
2. Machine snippet (merge into the schema positions):

math:

statements: ["T_arr = ( ∫ ( n_eff / c_ref ) d ell )"]

path: "gamma(ell)"

measure: "d ell"

symbols:

- { name: "n_eff", unit: "1", dim: "1" }

- { name: "c_ref", unit: "m·s^-1", dim: "L T^-1" }

check_dim: true

evaluation:

gates_hard: ["gate_accuracy>=0.99@7d"]

gates_soft: ["unit_cost<=1.0x@30d"]

VIII. Validation Rules (regex/consistency, Mandatory)

• References: ^EFT\.WP\.[\w\.]+ v\d+\.\d+(?:\.\d+)?:[PSMI]\d+(?:-\d+)?$.
• Equations: inline with backticks; if T_arr appears, math.path and math.measure must not be empty.
• Gates: gate_<metric>(>=|<=)[^@]+@[^ ]+$; statistics method and ci are required.
• Dimensions: math.check_dim=true; math.symbols[*].unit/dim required.

IX. Citation & Cross-Reference Style (Mandatory)