Skip to content

Simplify docs scaffolding #1433

Closed as not planned
4 of 4 issues completed
Closed as not planned

Description

@pelikhan

Organize documentation to gradually introduce complexity.

/scout

Activity

  1. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Agentic Issue Classifier triggered by this issue

  2. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Agentic Scout triggered by this issue

  3. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    🔍 Scout Research Report

    Triggered by @pelikhan on issue #1433

    Executive Summary

    Research shows that simplifying documentation scaffolding and organizing content to gradually introduce complexity is best achieved through progressive disclosure, layered architecture, and frameworks like Diátaxis. The current gh-aw documentation uses Astro Starlight with a reasonable structure, but can be significantly improved by:

    1. Adopting the Diátaxis framework (tutorials, how-to guides, reference, explanation)
    2. Implementing stronger progressive disclosure patterns
    3. Reducing initial cognitive load through better information architecture
    4. Creating clearer pathways from simple to complex concepts
    Click to expand detailed findings

    Research Findings

    The Diátaxis Framework

    The Diátaxis framework is the industry-standard approach for organizing technical documentation. It divides documentation into four distinct quadrants based on user needs:

    • Tutorials: Learning-oriented, takes beginners by the hand through a series of steps to complete a project
    • How-to guides: Task-oriented, shows how to solve specific problems
    • Reference: Information-oriented, technical descriptions of machinery and how to operate it
    • Explanation: Understanding-oriented, clarifies and illuminates particular topics

    Why it matters for gh-aw:

    • Major companies (Gatsby, Cloudflare, Vonage, Replit) have successfully reorganized their docs using Diátaxis
    • Replit scaled to $100M ARR while using Diátaxis to support diverse users navigating AI-driven development
    • The framework solves problems related to documentation content (what to write), style (how to write it), and architecture (how to organise it)

    Source: Diátaxis Documentation Framework

    Progressive Disclosure Principles

    Microsoft's documentation architecture research identifies progressive disclosure as a core design strategy:

    "Progressive disclosure keeps users focused on primary tasks and goals by minimizing distractions from secondary or irrelevant information and functionality. Provide the right amount of information and functionality at the right time."

    Key techniques:

    • Hiding, collapsing, expanding content as needed
    • Wizard patterns that break complex processes into sequential steps
    • Avoiding cognitive overload by revealing advanced features only when needed
    • Contextual help that clarifies without cluttering

    Application to gh-aw:

    • Start users with minimal YAML frontmatter examples
    • Introduce safe-outputs, tools, and network permissions progressively
    • Use expandable sections for advanced configuration options
    • Create clear pathways from "hello world" to complex multi-tool workflows

    Source: Microsoft - Recommendations for implementing an information architecture

    Layered Documentation Architecture

    Developer documentation best practices emphasize layered architecture that separates concerns:

    Three-tier structure:

    1. Business/Overview Layer: High-level value propositions, use cases
    2. Task/Guide Layer: Step-by-step instructions for common workflows
    3. Technical/Reference Layer: Detailed API specifications, parameters, options

    Navigation principles:

    • Users should be able to link between layers (jump from overview to technical details)
    • Each layer should have consistent information patterns across topics
    • Business users can stay at overview level; developers can dive deep as needed

    Current gh-aw structure analysis:

    • ✅ Has "Start Here" (entry layer)
    • ✅ Has "Workflows" reference documentation
    • ⚠️ "Guides" and "Application Areas" overlap in purpose
    • ⚠️ No clear tutorial/how-to separation
    • ⚠️ Reference documentation mixes concepts and specs

    Source: Document360 - API Developer Portal Best Practices

    Cognitive Load Reduction Strategies

    Research on developer experience and onboarding emphasizes:

    Information scent - Clear cues (link titles, summaries) that suggest where information can be found:

    • Current sidebar labels like "Workflows" are generic
    • Better: "Write Your First Workflow", "Workflow Configuration Reference"

    Minimize context switching:

    • In-context examples reduce need to jump between pages
    • Code snippets should be complete and runnable
    • Link to related concepts inline, but don't force navigation

    Onboarding best practices:

    • Single success path: One clear "happy path" to first success
    • Copy-paste ready examples: Reduce friction by providing working code
    • Progressive complexity: Start with minimal config, add features incrementally
    • Visual markers: Use callouts, tips, warnings to guide attention

    Sources:

    Examples of Effective Documentation Simplification

    Cloudflare developers docs (cited in Diátaxis testimonials):

    "While redesigning the Cloudflare developer docs, Diátaxis became our north star for information architecture. When we weren't sure where a new piece of content should fit in, we'd consult the framework. Our documentation is now clearer than it's ever been, both for readers and contributors."

    Gatsby reorganization:

    "By restructuring our documentation around the Diátaxis framework, we made it easier for users to discover the resources that they need when they need them."

    Recommendations

    1. Adopt Diátaxis Structure

    Reorganize sidebar into four clear sections:

    📚 Documentation
    ├── 🎓 Tutorials (Learning-oriented)
    │   ├── Your First Agentic Workflow
    │   ├── Adding Tools and Permissions  
    │   ├── Working with Safe Outputs
    │   └── Building a ChatOps Bot
    ├── 📋 How-To Guides (Task-oriented)
    │   ├── Configure Network Permissions
    │   ├── Debug Workflow Failures
    │   ├── Use MCP Servers
    │   └── Implement Custom Engines
    ├── 📖 Reference (Information-oriented)
    │   ├── Frontmatter Schema
    │   ├── Available Tools
    │   ├── Safe Outputs Reference
    │   └── CLI Commands
    └── 💡 Explanation (Understanding-oriented)
        ├── How Agentic Workflows Work
        ├── Security Model
        ├── Compilation Process
        └── When to Use Which Engine
    

    2. Create a Clearer Onboarding Path

    Implement a stepped learning journey:

    Level 1 - Quick Win (5 minutes):

    • Single file: issue-greeter.md with minimal frontmatter
    • No tools, no permissions beyond read
    • Just natural language + trigger

    Level 2 - Safe Outputs (15 minutes):

    • Add safe-outputs.add-comment
    • Show how AI output becomes GitHub comment
    • Explain permission separation

    Level 3 - Tools (30 minutes):

    • Introduce GitHub tools
    • Web search capabilities
    • File editing

    Level 4 - Advanced (1+ hours):

    • MCP servers
    • Custom engines
    • Network permissions
    • Complex workflows

    3. Progressive Disclosure in Reference Docs

    Frontmatter documentation should use expandable sections:

    ## Core Fields (Start Here)
    - `on:` - Workflow triggers
    - `permissions:` - Token permissions  
    - `engine:` - AI processor
    
    <details>
    <summary>Advanced Frontmatter Fields</summary>
    
    - `tools:` - Tool configuration
    - `network:` - Network access control
    - `cache:` - Dependency caching
    - `safe-outputs:` - Output processing
    
    </details>

    4. Improve Information Scent

    Replace generic labels with task-oriented ones:

    Current → Improved:

    • "Workflows" → "Workflow Reference"
    • "Tools" → "CLI Commands & Tools"
    • "Guides" → "How-To Guides"
    • "Application Areas" → "Use Case Examples"
    • "Start Here" → "Getting Started"

    5. Visual Documentation Patterns

    Add visual elements to reduce cognitive load:

    • Workflow diagrams: Show compilation process visually
    • Permission matrices: Table showing which safe-outputs need which permissions
    • Decision trees: "Which engine should I use?" flowchart
    • Annotated examples: Code snippets with inline explanations

    Key Sources

    1. Diátaxis Framework - Systematic approach to technical documentation authoring
    2. Microsoft - Information Architecture - Progressive disclosure and IA best practices
    3. Document360 - API Developer Portals - Layered architecture patterns
    4. ProprofS - Developer Documentation - Onboarding and DX principles
    5. Cloudflare, Gatsby, Replit case studies on Diátaxis adoption

    Suggested Next Steps

    1. Audit current documentation against Diátaxis quadrants

      • Tag each existing page as Tutorial/How-to/Reference/Explanation
      • Identify gaps (e.g., missing task-oriented guides)
    2. Create tutorial sequence

      • Draft "Your First Agentic Workflow" tutorial
      • Build progressive complexity ladder (5min → 15min → 30min → advanced)
    3. Reorganize sidebar in astro.config.mjs

      • Implement Diátaxis sections
      • Update labels for better information scent
    4. Add progressive disclosure

      • Use <details> tags for advanced content
      • Create "Simple" vs "Advanced" tabs where appropriate
    5. Test with new users

      • Run onboarding sessions with developers unfamiliar with gh-aw
      • Measure time-to-first-success
      • Iterate based on feedback

    AI generated by Scout

  4. pelikhan commented on Oct 10, 2025

    @pelikhan
    CollaboratorAuthor

    /plan look at the scout research report and generate a plan to rehaul the docs

    Agentic Plan Command triggered by this comment

  5. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1450

  6. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1451

  7. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1452

  8. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1453

  9. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1454

  10. pelikhan commented on Oct 10, 2025

    @pelikhan
    CollaboratorAuthor

    /plan look at the scout research report and generate a plan to rehaul the docs

    Agentic Plan Command triggered by this comment

  11. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1471

  12. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1472

  13. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1473

  14. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1474

  15. github-actions commented on Oct 10, 2025

    @github-actions
    Contributor

    Created related issue: #1475

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions