Skip to main content

How This Glossary is Built

This page documents the workflow, tools, and processes used to create and maintain the API Docs Glossary. Transparency about methodology helps contributors understand the project's standards and enables others to replicate similar documentation projects.

Four-stage glossary development workflow:


Research & Collection

Term identification happens through:

  • UW API Docs course materials, API documentation
  • industry standards such as RFC documents, OpenAPI Specification
  • authoritative sources like MDN Web Docs, official specifications
  • community needs identified through issues and discussions

Standardization

AI assistance, specifically Claude AI, standardizes entries to match the Style Guide requirements:

  • consistent formatting and structure
  • appropriate capitalization rules
  • proper anchor link syntax
  • related term connections

Review & Refinement

Human review ensures:

  • technical accuracy of definitions
  • source verification and proper citation
  • clarity and readability
  • appropriate examples for API documentation context

Testing & Validation

Each entry undergoes:

  • local build testing with npm start
  • anchor link verification
  • cross-reference validation
  • navigation panel review

AI Tools and Usage

Claude AI Role

Claude AI assists with:

TaskAI 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

What AI Generates

AI-assisted content:

  • initial term definitions based on provided sources
  • structure and formatting standardization
  • related term suggestions
  • example code blocks and tables

AI Limitations

AI assistance doesn't replace:

  • expert verification of technical details
  • judgment about term relevance
  • understanding of audience needs
  • decision-making about information architecture

Master Prompts

Reusable prompts maintain consistency across contributions.

Adding New Terms

Please generate glossary term entries for my API Docs Glossary:
https://github.com/rhyannonjoy/api-docs-glossary

Terms to add:
* [Term 1]
* [Term 2]
* [Term 3]

Please use [SOURCE NAME] as a source: [SOURCE URL]

I think these terms should go in [FILENAME], in the [SECTION NAME] section,
do you agree?

Here is my glossary style guide: [paste or link to style-guide.md]

Please generate glossary term entries in Markdown so that I can copy
them easily.

Standardizing Existing Entries

Please review these glossary entries against the style guide and 
suggest improvements:

[paste entries]

Style guide: [paste or link to style-guide.md]

Focus on:
- formatting consistency
- capitalization rules
- related term linking
- definition clarity

Quality Control Process

Verification Checklist

Before merging new entries:

  • 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 and 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 master prompts 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

Appropriate use cases:

  • formatting many entries consistently
  • generating table structures
  • drafting initial definitions from sources
  • finding anchor link references

Requires human judgment:

  • selecting which terms to include
  • verifying technical accuracy
  • choosing appropriate examples
  • determining term relationships

Maintaining Consistency

Style Guide Evolution

The Style Guide documents standards as they emerge:

  • new patterns get documented
  • edge cases get resolved
  • rules get refined based on experience

Version Control

Git history provides:

  • complete record of changes
  • reasoning through commit messages
  • ability to track when and why entries changed

Lessons Learned

What Works Well

AI assistance excels at:

  • applying consistent formatting across many entries
  • generating structured content from examples
  • suggesting related term connections
  • catching formatting inconsistencies

Human expertise essential for:

  • judging term relevance and priority
  • verifying technical accuracy
  • understanding audience needs
  • making editorial decisions

Continuous Improvement

The workflow evolves through:

  • contributor feedback
  • discovery of new patterns
  • refinement of master prompts
  • updates to tooling and automation

Replicating This Process

Others can adapt this workflow for similar documentation projects:

Prerequisites

  • clear style guide documenting standards
  • AI tool access - Claude AI, ChatGPT, etc.
  • version control system - Git
  • local testing capability
  • commitment to human review

Adaptation Tips

For different content types:

  • adjust master prompts to the domain
  • define what AI should and shouldn't do
  • document the specific quality standards

For different tools:

  • adjust master prompts to work with different AI assistants
  • adapt syntax for the build system
  • customize verification scripts

Key Success Factors

  1. Clear standards: comprehensive style guide
  2. Transparent process: document what AI does
  3. Human oversight: review everything
  4. Continuous refinement: improve prompts over time
  5. Testing automation: catch issues early

Feedback and Questions

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