Skip to main content

How This Glossary is Built

This page documents the workflow, tools, and processes used to create and maintain the ADG. Transparency about methodology helps contributors understand the project's standards and enables others to replicate similar documentation projects. Review the four-stage ADG development workflow:

Glossary development workflow in the style of a flowchart

Research & Collection

StageDescriptionKey Activities
Research & CollectionTerm identification from authoritative sourcesUW API Docs course materials; industry standards - RFC documents, OpenAPI Specification; authoritative sources - MDN Web Docs, official specifications; community needs via issues and discussions
StandardizationAI-assistance applies Style Guide formattingConsistent formatting and structure; appropriate capitalization rules; proper anchor link syntax; related term connections
Review & RefinementHuman review for accuracy and clarityTechnical accuracy of definitions; source verification and proper citation; clarity and readability; appropriate examples for API documentation context
Testing & ValidationContributors verify each entry before mergingLocal build testing with npm start; anchor link verification; cross-reference validation; navigation panel review

AI Tools & Usage

AI assists with content generation, standardization, code review, and feature development. While AI-generated content includes initial brainstorming and templating, AI-assistance doesn't replace expert verification, judgment on term relevance, understanding of audience needs, or decision-making about information architecture - and is susceptible to bias and hallucination, requiring explicit prompting for context-specific work.

AspectAI RoleHuman Role
Content GenerationDrafts initial entries from sourcesReviews, edits, verifies accuracy
StandardizationApplies style guide formattingEnsures consistency, fixes edge cases
Code ReviewSuggests improvementsMakes final decisions
Feature DevelopmentPair programming on toolingDirects implementation, tests

Prompt Templates

Reusable prompts maintain consistency across contributions. Review Prompt Templates; add more to expedite glossary workflows.


Quality Control Process

Verification Checklist:

  • Entry completion: definition in both category and Quick Reference
  • Technical accuracy: definition matches authoritative sources
  • Source verification: links work, citations are specific
  • Style compliance: follows all capitalization and formatting rules
  • Related terms: links point to existing entries only
  • Anchor links: all internal links tested and working
  • Examples: clear, relevant to API documentation context
  • Build success: npm run build completes without errors

Testing Commands:

# Verify build and check for broken links
npm run build

# Search for specific anchor references
grep -rn "](#anchor-name)" . --include="*.md"

# Test locally
npm start

Common Issues & Solutions:

IssueSolution
Broken anchor linksRun npm run build to identify, use grep to find references
Inconsistent capitalizationReview Capitalization section
Missing related termsCheck if linked terms have glossary entries
Vale warningsAdd ignore comments for intentional style guide breaks

Contribution Workflow

Contributors can use similar AI-assisted workflows:

  1. Create an issue describing the proposed additions or changes
  2. Use prompt templates to generate initial drafts with AI assistance
  3. Review and refine AI-generated content for accuracy
  4. Test locally with npm start before submitting
  5. Submit pull request with clear description of changes
  6. Reference the issue in the pull request

When to Use AI Assistance:

AspectUse AI ForUse Human Judgment For
EntriesFormatting many entries consistentlySelecting which terms to include
StructureGenerating table structuresDetermining term relationships
DefinitionsDrafting initial definitions from sourcesVerifying technical accuracy
LinksFinding anchor link referencesChoosing appropriate examples

Maintaining Consistency

The Style Guide documents standards as they emerge including new patterns, resolved edge cases, and rules refined based on experience. Version control using Git provides a complete record of changes, reasoning through commit messages, and the ability to track when and why entries have changed.


Lessons Learned

The Style Guide evolves to document new patterns, resolve edge cases, and refine ADG rules. Git history provides a complete record of changes and reasoning. The workflow itself improves through contributor feedback, discovery of new patterns, refinement of prompt templates, and updates to tooling. Adapt this workflow for similar docs projects:

AspectNeedsAdjustments
StandardsClear style guide documenting standardsAdapt prompt templates to the domain
AI ToolingAI tool accessEdit prompts for different AI assistants
Version ControlVersion control systemCustomize verification scripts
Build SystemLocal testing capabilityAdapt syntax for the build system
ProcessCommitment to human reviewDocument what AI should/shouldn't do

Key success factors include clear standards and a transparent process that includes human oversight, testing automation, and continuous refinement -

AspectAI Excels AtHuman Expertise Essential For
FormattingApplying consistency across all entriesJudging term relevance and priority
ContentGenerating structured content from examplesVerifying technical accuracy
ConnectionsSuggesting related term connectionsUnderstanding audience needs
QualityCatching formatting inconsistenciesMaking editorial decisions

Feedback and Questions

This workflow documentation evolves based on contributor experience and feedback. Questions about this workflow or suggestions for improvement? Open an issue.