Skip to content

System Overview

The site is no longer just a polished landing page. It is a system architecture whitepaper that explains how repository assets become an evidence-driven publication surface.

Five-Layer System

A five-layer system turns the repository into a review-grade publication surface

Each layer has one explicit job so the corpus, the site shell, and the GitHub Pages output remain aligned.

1Asset Layerrules/ stays canonical and defines the evidence boundary.2Snapshot Layersync-site-facts.mjs produces categoryDistribution, coverageBuckets, and qualitySignals.3IA LayerIntroduction, architecture, algorithms, performance, references, and evidence form the reading order.4Presentation LayerSignalGrid, TopologyMap, PipelineCanvas, and DossierMatrix expose metrics and arguments.5Publishing LayerVitePress and GitHub Pages publish the final whitepaper shell and its publish evidence.

The five-layer system

1. Asset layer

  • Job: keep the rule corpus canonical.
  • Location: rules/
  • Why it matters: every metric, evidence page, and claim must be traceable to repository state.

2. Snapshot layer

  • Job: translate repository state into structured data.
  • Location: docs/scripts/sync-site-facts.mjs, docs/.vitepress/site/facts.ts
  • Why it matters: the snapshot layer keeps metrics, diagrams, and dossiers consistent instead of hand-maintained.

3. IA layer

  • Job: decide the shortest review path for senior readers.
  • Location: docs/.vitepress/site/atlas.ts
  • Why it matters: the site now explains context first, then architecture, then the rule orchestration flow, and only then the evidence corpus.

4. Presentation layer

  • Job: render a coherent visual language for topology, pipeline, evidence, and references.
  • Location: docs/.vitepress/theme/
  • Why it matters: figure tokens remove the old light/dark SVG fragility and make the diagrams reusable across chapters.

5. Publishing layer

  • Job: publish the final reading experience.
  • Location: docs/.vitepress/config.ts
  • Why it matters: the publishing layer keeps structure, metadata, and GitHub Pages output aligned without inventing new facts.

Engineering decisions

Repository assets outrank page copy

The rule corpus is not supporting material. It is the upstream source that drives metrics, diagrams, and the evidence layer.

The snapshot layer outranks static metrics

The site now carries categoryDistribution, coverageBuckets, qualitySignals, and timelineSignals. That lets the architecture, performance, and references chapters share one factual base.

Rule orchestration must be explainable

Instead of stopping at a category index, the site documents the rule orchestration logic that turns isolated assets into review paths, composition patterns, and publish evidence.

Continue reading

  1. Algorithms Overview
  2. Performance Analysis
  3. References

Reframing 132+ rule assets across 32+ domains as a whitepaper, algorithm dossier, and evidence library