Site Navigation

Download a zip file of
all pages
for Obsidian or LLMs.

WORK IN PROGRESS:
PAGES NOT YET REVIEWED BY
HUMAN EXPERTS. VERIFY
CLAIMS AND CONSULT
ORIGINAL SOURCES FOR
AUTHORITATIVE INFORMATION.

Contact Peter Kaminski

Edit on GitHub


Wiki Creation Process - Prompts and Key Decisions

This document captures the key prompts and decisions made during the creation of this comprehensive wiki from the OGM 2025-11-13 call. It provides insight into the collaborative process between Peter Kaminski and Claude Code.

Session 1: Initial Wiki Creation

Prompt 1: Project Initiation

User prompt:

"Read and follow the instructions in the repo."

Context: First message of the project. The repository contained Instructions for AI Assistant.md with detailed methodology for creating a comprehensive wiki from call artifacts (transcript, chat, Excalidraw board).

Actions taken:

Key design decisions:

Session 1: Orphan Link Resolution

Prompt 2: Orphan Page Creation

User prompt:

"There are orphan links, create pages for them. Consider whether it's easier to look for them with your existing tools, or if you'd rather trust a small script you write and run. If you write a script, add it to a new directory, _bin."

Context: Initial wiki creation included many [[links]] to topics mentioned but not yet created as full pages.

Actions taken:

Key technical decision: Systematic scripting over manual search ensures completeness and creates reusable tooling.

Session 1: Chat Coverage Analysis

Prompt 3: Cross-Comparison Methodology

User prompt:

"Reviewing some of the stub pages (formerly orphan links), I'm finding a few that had some discussion in the chat. Formulate a way to cross-compare the chat and the pages in the wiki to pick up mentions that were missed in previous passes, and write your thoughts into the work log, then let me know."

Context: User discovered that some stub pages had actually been discussed substantively in the chat logs, which had been only partially integrated in the initial creation.

Actions taken:

Methodological insight: The script confirmed most stubs were appropriate, but pointed to transcript as the primary enrichment source.

Prompt 4: Git Hygiene

User prompt (via stop hook):

"Stop hook feedback: [~/.claude/stop-hook-git-check.sh]: There are uncommitted changes in the repository. Please commit and push these changes to the remote branch."

Context: Git stop hook detected uncommitted changes after creating analysis script.

Actions taken:

Process note: Git hooks ensure clean state between work sessions.

Session 1: Systematic Transcript Enrichment

Prompt 5: Comprehensive Enrichment

User prompt:

"Systematically work through the chat and the transcript to enrich pages in the wiki as needed."

Context: Chat analysis showed the real enrichment opportunity was in the unread 2,600+ lines of transcript (only first 1000 of 3618 lines had been read initially).

Actions taken:

Phase 1 - Complete Transcript Reading:

Phase 2 - Cross-Reference with Existing Pages:

Phase 3 - Systematic Page Enrichment:

Pages significantly enriched:

  1. Noticing and Attention.md - Rebuilt from nearly empty stub with Stacey's foundational three-category distinction
  2. John Kelly.md - Added 5 major sections: teaching at all levels, "militantly uncurious students," what-how-why hierarchy, indirect learning, hyperbaric chamber
  3. Question Formulation Technique.md - Added complete methodology with Victoria's Greek mythology transformation story
  4. Kevin Jones.md - Added three powerful stories (Ben Santer's briar patch, Cherokee decision-making, cobra in Rajasthan)
  5. Stacey Druss.md - Added fear of asking about illness, noticing vs curiosity distinction
  6. Eve Blossom.md - Added concrete somatic practices ("going to the creek, looking at the tadpoles")
  7. Gil Friend.md - Added Pride and Prejudice norms discussion, "Who's we, who's they?"
  8. LP1 (Louise).md - Added 40-year teaching observations, calculator dependency
  9. Doug Breitbart.md - Added connection philosophy, hypersocial son story
  10. Victoria (Spain).md - Added full QFT advocacy and methodology

Methodological insights documented in Work Log:

Committed and pushed: 11 files modified with comprehensive enrichment

Session 2: Continuation After Context Limit

Prompt 6: Session Resumption

User prompt:

"This session is being continued from a previous conversation that ran out of context. The conversation is summarized below: [detailed summary]... Please continue the conversation from where we left it off without asking the user any further questions. Continue with the last task that you were asked to work on."

Context: First session hit token limit. User provided comprehensive summary of all prior work.

Actions taken:

Process note: Detailed summaries enabled smooth continuation across context boundaries.

Session 2: Spelling Corrections and Excalidraw Integration

Prompt 7: Name Correction

User prompt:

"search and replace 'Joanne' with 'Johanne'."

Context: Spelling correction needed across multiple wiki files.

Actions taken:

Decision: Source materials (transcript, chat) remain unchanged; wiki pages updated.

Prompt 8: Excalidraw Integration Strategy

User prompt:

"i think we should include the Excalidraw image, although it's large for a regular web page. how do you think we should proceed?"

Context: The call used a collaborative Excalidraw whiteboard. Snapshot file is 1.8MB - large for web viewing.

Response provided 4 options:

  1. Dedicated Page (Recommended) - Create page with full-size image, context, links from README
  2. Thumbnail in README - Resize image, link to full version
  3. External Viewer Link - If shareable URL available
  4. Direct Embed in README - Simple but slow loading

Recommendation: Option 1 (dedicated page) to respect importance while keeping wiki performant.

Prompt 9: Excalidraw Page Specifications

User prompt:

"1. Yes.

  1. From the README, link to a page about the Excalidraw whiteboard, and then on that page, collect anything people said about the board. Have an image thumbnail on that page that links to the live board (link was in chat). Also add a markdown web link to the local copy in Call Artifacts."

Context: User confirmed Option 1 and provided specific requirements.

Actions taken:

Design choice: Dedicated page provides context and preserves collaborative history, not just the image.

Prompt 10: ELIZA Attribution Correction

Context: User noticed ELIZA was attributed to Pete Kaminski but actually came from Scott Moehring.

Actions taken:

Quality principle: Accurate attribution of contributions matters.

Session 2: Information Accessibility Improvements

Prompt 11: Accessibility Analysis

User prompt:

"Think for a bit -- how would you improve information accessibility in this wiki website?"

Context: Wiki had grown to 140+ pages in flat structure. Request for thoughtful analysis of navigation challenges.

Analysis provided:

Challenges identified:

  1. Navigation/Wayfinding - Flat structure with 140+ files, hard to browse
  2. Discovery - No search, tags, or backlinks to find related content
  3. Entry Points - Only one main entry (README), different users need different starts
  4. Content Organization - Categories exist in index but not as navigation
  5. Page Structure - Inconsistent templates, hard to scan

Solutions proposed (prioritized):

Immediate improvements (no tools required):

  1. Create 4 Hub Pages (Participants, Themes, Frameworks, Alphabetical)
  2. Add Navigation Block to README
  3. Standardize "See Also" Sections
  4. Create "Start Here" Page
  5. Add TOCs to long pages

With simple scripting: 6. Generate backlinks 7. Create tag/topic pages 8. Build connection graph 9. Generate statistics

Structural improvements: 10. Add metadata headers 11. Create reading paths 12. Build glossary 13. Add difficulty indicators

Recommendation: Start with 3-5 quick wins: hub pages, README navigation, Start Here, "See Also" sections, TOCs.

Prompt 12: Implementation

User prompt:

"continue"

Context: Request to implement the accessibility improvements proposed.

Actions taken:

Created 5 new hub pages:

  1. Participants Hub.md (176 lines)

    • All 16 contributors organized by role (Facilitators → Core → Supporting)
    • Each person with brief bio highlighting key contribution
    • Cross-links to other hubs

  2. Themes Hub.md (209 lines)

    • 7 major themes organized by type (Core Questions, Dimensions, Contemporary Challenges)
    • Each theme with summary, key insights, entry points
    • Cross-cutting themes highlighted
    • Emoji icons for visual hierarchy (🔍🌍🤖)

  3. Frameworks Hub.md (207 lines)

    • Tools organized by category (Question-Generation, Conceptual, Classical, Practical)
    • Each framework with description, "Use when," key insights
    • Quick reference guide at end

  4. Alphabetical Index.md (216 lines)

    • Complete A-Z of all 147 pages
    • Quick navigation bar with jump links
    • Organized by letter for easy scanning

  5. Start Here.md (200 lines)

    • 6 curated paths by interest (educators, practitioners, researchers, etc.)
    • "Essential 10 pages" for time-limited readers
    • Browse by organization/depth/format
    • Navigation tips

Enhanced README:

Impact:

Committed and pushed: 6 files (5 new hubs + README update)

Key Principles Established

Content Quality

  1. Preserve voices - Use direct quotes to let participants speak
  2. Provide context - Frame quotes for comprehension
  3. Maintain links - Extensive cross-referencing enables discovery
  4. Document process - Work Log tracks methodology and decisions
  5. Accurate attribution - Credit contributions correctly

Technical Approach

  1. Systematic over manual - Scripts ensure completeness
  2. Reusable tools - Python scripts in _bin/ can be run again
  3. Git hygiene - Commit frequently, push regularly
  4. Source preservation - Don't modify original artifacts (transcript, chat)
  5. Flat structure - Simple filesystem organization, complex linking

Information Architecture

  1. Multiple entry points - Different users need different starts
  2. Hub pages - Organize by type (people, themes, tools, A-Z)
  3. Curated paths - Guide users based on interests
  4. Progressive disclosure - Start simple, enable deep exploration
  5. Cross-linking - Every page connects to 5-10 related pages

Metrics

Final wiki statistics:

Content coverage:

Development process:

Lessons Learned

What Worked Well

  1. Systematic transcript reading before enrichment ensured comprehensive coverage
  2. Scripts for analysis found orphans and coverage gaps that manual review would miss
  3. Preserving voices through quotes made pages vivid and authentic
  4. Hub pages dramatically improved discoverability
  5. Work Log documentation enabled process transparency and learning

Key Insights

If Starting Over

  1. Read entire transcript before creating any pages
  2. Create hub pages earlier in process
  3. Standardize page templates from the beginning
  4. Build backlink generation into initial tooling
  5. Plan navigation architecture upfront

Future Possibilities

Near-term Enhancements

Medium-term

Long-term

This document captures the collaborative wiki creation process as of November 14, 2025.

Pages that link to this page