A deep dive into the Sound, Space, and Time dimensions of the Cormorant Foraging Framework.
The framework maps three observable physical dimensions to information processing paradigms:
| Dimension | Physical Basis | Scoring Method | Key Property |
|---|---|---|---|
| Sound | Acoustic superposition | Additive | Non-reversible propagation |
| Space | Cartesian coordinates | Multiplicative | Static snapshots |
| Time | Temporal flow | Exponential decay | Irreversible history |
Each dimension emerged independently through production systems before the underlying pattern was recognized.
Sound waves combine through superposition—multiple acoustic signals add together linearly. When two sound waves meet, their amplitudes sum at each point in space and time. This additive property maps directly to urgency assessment.
Key Properties:
- Linear combination - Effects sum independently
- Non-reversible - Sound propagates forward in time
- Attention-grabbing - Louder signals dominate awareness
Score = Σ(Factor_i × Weight_i) - Σ(Risk_j × Weight_j)
Where:
- Factors contribute positively (opportunity signals)
- Risks contribute negatively (warning signals)
- Weights modulate importance
Domain: Fantasy sports player intelligence
Formula:
ChirpScore = (RecentPerformance × 0.4) +
(OpportunityScore × 0.3) +
(ScheduleDifficulty × 0.2) +
(VolumeTrend × 0.1) -
(InjuryRisk × 0.3)
Performance:
- 78% accuracy predicting breakout players
- Processes real-time game data
- Deployed as MCP server
Observable Anchoring:
- Recent stats (last 3 games)
- Opponent defensive rankings
- Injury reports
- Target share percentages
Use additive scoring when:
- Partial information has value - Incomplete data still provides signal
- Speed matters - Fast decisions needed
- Multiple independent factors - No single factor should block action
- Communication urgency - Alerting, warnings, notifications
Examples:
- Risk assessment dashboards
- Alert prioritization systems
- Real-time opportunity detection
- Urgency scoring for support tickets
Three-dimensional space uses Cartesian coordinates (x, y, z) where each dimension is independent but multiplicative in volume calculation. Volume = length × width × height. If any dimension is zero, the entire volume collapses to zero.
Key Properties:
- Multiplicative relationships - Dimensions scale each other
- Zero collapse - Missing one dimension eliminates the whole
- Static snapshots - Spatial relationships at a point in time
- Structural mapping - Reveals organization and hierarchy
Score = Factor_1 × Factor_2 × Factor_3 × ... × Factor_n
Where:
- Each factor represents an independent dimension
- Zero in any factor reduces score to zero
- Higher values in any dimension amplify the total
Domain: Database schema intelligence
Formula:
ICE Score = (Insight × Context × Execution) / 100
Where:
- Insight (0-10): Value of optimizing this query/schema
- Context (0-10): Understanding of relationships and dependencies
- Execution (0-10): Feasibility and effort estimate
Performance:
- 398 passing automated tests
- Analyzes Cloudflare D1 databases
- Identifies optimization opportunities
Observable Anchoring:
- Table cardinalities (row counts)
- Foreign key relationships
- Index presence/absence
- Query patterns in schema
Key Insight: If Context = 0 (don't understand the schema), the entire ICE score becomes 0 regardless of Insight or Execution values. This prevents action without understanding.
Use multiplicative scoring when:
- All factors required - Missing any dimension should block action
- Structural analysis - Mapping relationships and hierarchies
- Interdependencies matter - Factors amplify or nullify each other
- Snapshot assessment - Static evaluation at a point in time
Examples:
- Prioritization frameworks (ICE, RICE)
- Database optimization
- Architecture analysis
- Relationship mapping
Time flows unidirectionally from past to future, creating an irreversible history. Information relevance naturally degrades over time like a boat's wake that gradually dissipates. Recent events have stronger impact than distant ones.
Key Properties:
- Exponential decay - Relevance decreases continuously
- Recency bias - Recent information weighted higher
- Irreversibility - Cannot undo the passage of time
- Memory persistence - Some signals last longer than others
Relevance = BaseRelevance × e^(-age/halfLife) × AccessBoost
Where:
BaseRelevance: Initial importance of the informationage: Time since creation/last modificationhalfLife: Time period where relevance drops to 50%AccessBoost: Multiplier for recently accessed items
Domain: Context continuity management
Formula:
ContextRelevance = InitialScore × e^(-days/7) × (1 + 0.5*recentAccess)
Performance:
- 85% context maintenance across sessions
- Manages conversation history
- Temporal context protocol
Observable Anchoring:
- File modification timestamps
- Last access times
- Creation dates
- Logged event sequences
Key Insight: Information doesn't disappear—it fades. A file modified 30 days ago has lower relevance than one modified yesterday, but a 30-day-old file accessed today gets a boost.
Use exponential decay when:
- Recency matters - Recent information more valuable
- Memory management - Limited capacity requires prioritization
- Continuity tracking - Maintaining context over time
- Historical patterns - Understanding trends and sequences
Examples:
- Cache eviction policies
- Session management
- Document relevance ranking
- Activity feed prioritization
- Log analysis
Most real-world systems combine dimensions:
Example: Task Prioritization System
Priority = (Urgency + Impact) × Feasibility × e^(-age/halfLife)
↑ ↑ ↑
Sound Space Time
- Sound (Urgency + Impact): Additive factors for immediate attention
- Space (× Feasibility): Multiplicative gate (can't do impossible tasks)
- Time (decay): Older tasks lose relevance
The framework intentionally keeps dimensions pure in their implementations:
- ChirpIQX: Only additive scoring
- PerchIQX: Only multiplicative scoring
- WakeIQX: Only exponential decay
This purity makes each system easier to reason about and composable.
Every dimension ties to measurable, observable properties:
- Performance metrics (percentages, counts)
- Difficulty ratings (opponent strength)
- Risk indicators (injury status)
- Volume measurements (usage rates)
- Structural properties (table sizes)
- Relationships (foreign keys)
- Presence/absence (indexes, constraints)
- Cardinalities (one-to-many, many-to-many)
- Timestamps (creation, modification)
- Access logs (last opened, viewed)
- Event sequences (ordered history)
- Age calculations (duration since event)
No speculation. No predictions. Only observable signals.
- Intent - No mind-reading or assumption about purpose
- Normative judgments - No "should" or "ought"
- Predictions - No forecasting future states
- Hidden information - Only what's measurable
These constraints ensure the framework stays grounded in physics rather than drifting into speculation.
- Commutative: A + B = B + A
- Associative: (A + B) + C = A + (B + C)
- Identity: 0 is identity element
- Monotonic: Adding positive values always increases score
- Commutative: A × B = B × A
- Associative: (A × B) × C = A × (B × C)
- Identity: 1 is identity element
- Zero annihilator: 0 × anything = 0
- Continuous decay: No sudden drops
- Half-life: Predictable degradation rate
- Non-negative: Relevance never goes negative
- Asymptotic: Approaches but never reaches zero
- Derived Layers - How DRIFT and Fetch build on these dimensions
- Applications - Real-world implementations
- Empirical Validation - Performance data
"Three dimensions. Three metaphors. Three production systems." 🐦