Skip to content
Doppl-LifePublic

About

Initial aGarden markdown dataset for Doppl judgment and kernel node work

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

36 Commits

Folders and files

Repository files navigation

aGarden

A living showcase of the Doppl method — an Obsidian vault of real runs, where each idea is grown from a case study into a recovered problem and then into one or more actionable answers. This is the floor model: the show, and also a thing that stands on its own.

The idea

Most analysis stops at the surface complaint. Doppl keeps going. It takes a case study — a situation worth understanding — recovers the actual problem hiding under the obvious one, then breeds that into doppls: the unlocks, opportunities, or solutions that follow once you see the real constraint.

The durable artifact is plain markdown. A human can read a node directly; a machine can parse it into a typed shape. Nothing here is a black box — every conclusion carries its lineage, its evidence, and its score.

The spine

case_study → problem_recovery → doppl → (your move)
  • case_study — the seed. A situation, stated plainly. Not yet a claim, so it isn't scored.
  • problem_recovery — the recovered problem: surface complaint → deleted assumption → hidden variable → actual problem. Scored.
  • doppl — the leaf: a claim, its implications, and the opportunities it opens. A single problem can grow several distinct doppls. After a doppl, the path points out of the system — into your action.

Each arrow is one pass of the engine: it generates candidate ideas, measures them on novelty and grounding, selects survivors, checks feasibility, and a judge rates the result. The winner is compiled into the next node.

What's in this vault

Path What it holds
flow/ The node graph as a nested lineage tree. Each node is one markdown file (## Trace · ## Discovery · ## Growth · ## Path); children nest under the node they grew from.
stock/ Durable domain memory. Admitted discoveries grouped by field — each a load-bearing fact with a synopsis, a grounding line, and a deep-linkable anchor. Stages read stock before searching outward again.
ratings-ledger.json The human-ratings ledger: per scored node, the {rater, score} entries that average into that node's scores.human / scores.n. Empty until people rate — the node files hold everything else.
deprecated/ Retired export dumps from earlier kernel versions. Kept for reference; not the source of truth.

Reading it

Open the folder as an Obsidian vault. The graph view shows the lineage of every idea and the links from nodes into the stock fields they drew on. Every cross-document reference in this vault is an Obsidian-compatible wikilink — [[slug-id]] for a file, [[slug-id#^anchor]] for a specific block — so the graph resolves end-to-end: a node's prev_id points at its parent node, and each Discovery finding deep-links the exact stock load-bearing fact it used.

Or just read the markdown. Every node opens with its one-line headline; the body walks from inherited context (Trace) through what was found (Discovery) to what was concluded (Growth), ending with where the flow points next (Path).

How scoring reads

  • Measurements (0…1) — instrument readings. Novelty = how far the idea reaches past what's already in the record; grounding = how well it lands on something true and testable.
  • Ratings (−5…+5) — judgments of worth. The judge fills five axes (Novelty, Grounding, Falsifiability, Cost-efficiency, Relevance) and boils them to scores.judge. A human gives one gut-read slider (scores.human). Negative means value-subtracting, not merely ineffective.

Vault config in git

Only shared meaning is versioned; personal view-state is not. Connections (wikilinks), node content, the ledgers, and the graph color scheme are signal and get committed. Your zoom level, pane layout, and open tabs are noise and must never compete across machines.

  • .obsidian/workspace.json (window layout, open tabs, active file, last-opened) is git-ignored. Each person keeps their own.
  • .obsidian/graph.json is tracked but frozen so the shared colorGroups stay in sync without your zoom (scale) creating churn. Each collaborator runs this once per clone:
git update-index --skip-worktree .obsidian/graph.json

To intentionally change the shared scheme (e.g. recolor a stage), unfreeze, edit, commit, then re-freeze:

git update-index --no-skip-worktree .obsidian/graph.json
#   ...edit colorGroups, git add, commit, push...
git update-index --skip-worktree .obsidian/graph.json

Anyone who pulls gets the new colors automatically; their local zoom is unaffected.

Where the model lives

The contracts and the model that produced this vault live in the doppl-prime repo (my-docs/the-hut/ and contracts/). aGarden is an instance of that model — but it's meant to be legible without it.

About

Initial aGarden markdown dataset for Doppl judgment and kernel node work

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors