Quartz Migration Findings

Updated: 2026-02-27

Status Snapshot

The migration from monolithic source files to Quartz-native pages is complete at the structural level.

What is now true:

  • The site uses a foldered note graph (methodology, framework, factors, chokepoints, investment-theses, predictions, fact-check, context-primer, glossary, appendices).
  • The split pages are standalone markdown pages (no runtime transclusion dependency on monolithic compendium files).
  • Cross-page navigation is hub-driven via index pages and wikilinks.
  • Deployment is now Cloudflare-first, with GitHub Actions used as CI and GitHub Pages retained as manual fallback only.

Archive Decision (Monoliths Hidden, Not Deleted)

We intentionally preserve the original monoliths for traceability and future audits.

  • Archived files:
    • content/archive/allthingsfinancial/compendium.md
    • content/archive/allthingsfinancial/compendium-appendices.md
  • Hidden from site generation via Quartz ignore pattern:
    • archive/**

This keeps source lineage available in git while removing monolithic pages from public navigation and output.

What Improved

  • Navigation is no longer blocked by very long documents.
  • The content graph is now decomposed into reusable topic pages.
  • Section-level ownership and future editing are simpler.
  • It is now practical to rewrite pages incrementally without touching a single giant file.

Remaining Editorial Debt

The migration was structurally successful, but many pages are still source-faithful transfers rather than fully Quartz-native editorial notes.

Current weaknesses:

  • Several pages remain too long and should be split by claim cluster.
  • Callout idioms are inconsistent (some pages use plain prose where claim/assessment callouts would improve scanability).
  • Verification semantics (VERIFIED, MISLEADING, FALSE, UNVERIFIED) are present but not yet normalized into repeatable block patterns.
  • Some pages still carry report-like prose density instead of note-style synthesis.

Quartz-Native Rewrite Standard (Phase 2)

Each analytical page should converge toward this pattern:

  1. What this page covers (2-4 sentence framing)
  2. Channel claims as [!quote] callouts
  3. Independent assessment as [!success], [!warning], or [!danger] callouts
  4. Evidence table (claim, verdict, source type)
  5. What would falsify this section
  6. Related notes with 3-8 high-value wikilinks

Suggested verdict callout convention:

> [!quote] Channel Claim
> 
> ...
 
> [!success] Independent Assessment (VERIFIED)
> 
> ...

Use parallel variants for other verdict classes:

  • [!warning] Independent Assessment (MISLEADING)
  • [!danger] Independent Assessment (FALSE)
  • [!note] Independent Assessment (UNVERIFIED)

Rewrite Backlog (Priority Order)

Priority 1: High-impact investor pages

  • investment-theses/*
  • predictions/*
  • fact-check/*

Reason: these drive most practical usage and benefit most from tighter evidence structure.

Priority 2: Framework coherence pages

  • framework/*
  • chokepoints/*

Reason: these define the logic model and improve consistency across downstream theses.

Priority 3: Reference and onboarding pages

  • context-primer/*
  • glossary/*
  • appendices/*

Reason: structurally adequate already; mostly editorial polish and cross-link enhancement.

Definition of Done for Phase 2

Phase 2 is complete when:

  • Every page has explicit claim-vs-assessment separation.
  • Every investment/prediction page includes falsification criteria.
  • Fact-check pages use consistent verdict callouts and table schema.
  • No page includes stale references to archived monolith paths.
  • Hubs link to concise notes rather than long report-like blocks.

Operational Notes

  • Use QUARTZ_BASE_URL in CI and deployment environments.
  • Keep upstream pointed at jackyzha0/quartz for future sync.
  • Keep origin pointed at full-automagic/macronomicon for publishing.