Note Split (Atomization)
Goal
Turn one long, fuzzy document into multiple notes that can stand alone and connect cleanly.
Quick start
- Identify 3–7 candidate atomic concepts.
- For each new note: one-sentence core + original source location + ≥2 link suggestions.
Workflow
Step 1: Concept identification (three layers)
- Surface concepts (explicit terms)
- Implicit concepts (assumptions / premises)
- Core insight (what the author really claims)
Step 2: Boundary test
- If removing any part breaks the concept, it’s not atomic yet.
- If the concept can explain other phenomena, it’s a good atomic note.
Step 3: Split plan
Provide a mapping table:
- original section → new note title → core concept → note type (concept/method/case)
Step 4: Title patterns
- Definition: “What is X? core elements”
- Method: “Method X: key steps”
- Insight: “On [topic]: [insight]”
Step 5: Links
Include:
- sibling links (notes split from the same source)
- topical links (same theme)
- contrast links (alternative view / counterexample)
Output template
## 📋 Split Plan
### Source note
- **File**: [name]
- **Length**: [words/lines]
- **Concepts identified**: [N]
### Proposed new notes
#### 1) [New note title]
- **Core concept**: [one sentence]
- **Source location**: [section/paragraph]
- **Suggested links**: [[Link 1]] [[Link 2]]
### Keep-as-index suggestion
- Keep the original as an index note? yes/no (why)
Quality checks
- Each new note is atomic and independently understandable
- Each includes source location
- Clear linking between split notes and the wider network