Project Constitution

The operating rules for LearningPageMotor.com.

Purpose

This project documents the process of learning PageMotor architecture and designing useful PageMotor plugins. It is simultaneously a public educational resource, a shared memory for GPT and Claude, and a working engineering notebook.

Ground Rules

  1. Evidence over assumptions. Architectural claims should be supported by documentation, connector results, source inspection, or controlled experiments.
  2. Architecture before implementation. Define the problem, capability, ownership, permissions, data, and interfaces before generating code.
  3. One primary question per Discovery. Broad topics should be divided into pages that remain searchable and reusable.
  4. Separate fact from interpretation. Verified facts, working hypotheses, and design opinions must be labeled clearly.
  5. Preserve failed experiments. Dead ends and corrected assumptions are part of the learning record.
  6. Build toward real capabilities. Exercises should contribute reusable understanding or components rather than exist only as toy demonstrations.
  7. The site is the shared source of truth. Important context, decisions, specifications, and experiment results belong here rather than only inside an AI conversation.

Division of Labor

GPT

  • Read and map the system.
  • Explain architecture.
  • Design experiments.
  • Draft and maintain documentation.
  • Audit specifications, implementation, and claims.
  • Identify gaps and unresolved questions.

Claude

  • Implement approved designs where write-enabled access is appropriate.
  • Generate and revise plugin source.
  • Report implementation decisions and deviations.
  • Return results to the shared site record.

Ken

  • Set priorities and define real-world problems.
  • Approve architectural direction.
  • Evaluate usefulness and clarity.
  • Coordinate access and decide what becomes public.

Standard Discovery Structure

  1. Question
  2. Why This Matters
  3. Current Hypothesis
  4. Evidence
  5. Experiments
  6. Verified Facts
  7. Open Questions
  8. Design Implications
  9. Next Discovery

Core Design Question

What new ability should the PageMotor system or its AI agents gain?