Skip to content

FCC for Technical Writers

Welcome, technical writer. This track is written for documentation professionals — staff technical writers, freelance doc consultants, content strategists, DX (developer experience) writers, and content-ops folks — who want to adopt FCC patterns for structured, persona-driven, diagram-rich documentation.

FCC's documentation philosophy and infrastructure are unusually mature for an OSS project its size. It combines hand-authored narrative with generated reference material; it treats Mermaid diagrams as first-class content; it uses the R.I.S.C.E.A.R. specification as a contract for role-based docs; and it publishes simultaneously to web, PDF, EPUB, DOCX, and slides via a single source.

What you will learn

  • A concrete Mermaid-diagram convention that produces publication-grade output across web, PDF, and EPUB.
  • How to drive documentation from persona YAML files so that role descriptions stay current.
  • A publication pipeline pattern (Quarto + Pandoc + Typst) that scales from blog post to book.
  • Glossary-governance techniques for multi-audience, multi-format documentation.

How this track is organized

Page Focus Effort
Getting Started Orientation; tool survey 30 min
Mermaid Conventions Figure naming, captions, rendering 1 hour
Persona-Driven Docs YAML → Markdown pipeline 2-3 hours
Publication Pipeline 101 Multi-format publishing 3-4 hours
Glossary Governance Terminology discipline 1-2 hours
Further Reading References and tools reference

Prerequisites

  • Comfort with Markdown.
  • Familiarity with Git (commits, branches, pull requests).
  • Experience with at least one docs pipeline (MkDocs, Docusaurus, Hugo, Read the Docs, etc.).
  • Willingness to edit YAML.

You do not need:

  • To write Python (but you will read some).
  • Extensive LaTeX experience (Quarto hides most of it).
  • Prior AI or ML background.

Why FCC is worth studying

Three things FCC does well:

  1. Diagrams as code. ~280 Mermaid blocks across the docs, rendered to SVG + PNG for HTML and PDF respectively, via a Pandoc Lua filter. Diffable in PRs. Reproducible.
  2. Generated reference docs. 173 model cards auto-generated from persona YAMLs. Zero drift between code and docs.
  3. Multi-format output. One source tree produces a website, a 4-book series, executive briefings, academic papers, and reference cards via a single make pub-all.

Each of these patterns is individually adoptable. You do not have to adopt all three to get value.

What you will not learn here

  • Copy-editing. There are better resources.
  • SEO or content-marketing strategy. Off-topic.
  • Structured-content standards (DITA, DocBook). FCC chose Markdown-plus-YAML instead; you will not find DITA tooling here.
  • Headless-CMS patterns. Different architecture.

A note on audience overlap

This track's neighbors in docs/tutorials/ include:

  • for-open-source-maintainers — you will find useful material there on docs-as-code-patterns.md and release-cadence-lessons.md.
  • for-students — if you write educational content, assignment-patterns translates to tutorial structure.
  • for-policy-makers — if your users include policy audiences, the ecosystem-glossary is a model of plain-language definitions.

Feel free to dip in; the tracks are independent but complementary.

The 15-minute orientation

Before you dive into this track's pages, spend 15 minutes browsing the FCC repo itself. Look at:

  • docs/index.md — the root entry point.
  • docs/decisions/ — the 5 ADRs.
  • docs/model-cards/ — one auto-generated card.
  • docs/guidebook/ — one chapter.
  • publications/ — the publishing directory.
  • src/fcc/templates/docs/ — the Jinja templates that generate the reference docs.

You will notice the separation between narrative (hand-authored) and reference (generated). That separation is the heart of the pattern.

Starter mental model

Think of FCC's docs as having three layers:

  1. Source layer: YAML personas, JSON scenarios, hand-authored Markdown narrative.
  2. Generation layer: Jinja templates + Pandoc + Quarto, which turn source into publishable Markdown.
  3. Rendering layer: MkDocs, Pandoc, Typst, producing web, PDF, EPUB, DOCX, PPTX.

Your job as an adopting writer is usually in the narrative source + the generation templates. You rarely touch rendering.

Where next

If you want to see quick wins, start with Mermaid Conventions — it is the lowest-friction pattern and most visible to readers. If your pain is docs drift, go to Persona-Driven Docs. If you need end-to-end understanding before adopting anything, Getting Started.

Related: ADR-0001 Plugin System, ADR-0002 Persona Schema, codename decoder.