Time-Travel

Navigate QNTX's attestation history to understand how knowledge evolved.

Core Concept

Every attestation has a timestamp (✦) and causal links (⟶), creating a complete navigable history of system intelligence.

What Makes It Unique

• Causal chains - See why things happened, not just what
• Live reconstruction - Rebuild exact system state at any moment
• Cross-domain - Attestations span code, data, and decisions
• AI-queryable - LLMs can understand and navigate the timeline

Primary Use Case

Watch knowledge evolution - how the system's understanding developed over time through continuous learning cycles.

Temporal Dimensions

• System Time - When events occurred in QNTX's overall timeline
• Stream Time - Position within a video/data stream (e.g., VidStream frame timestamps)
• Window Width - How much temporal context to show (affects graph filtering)
• Cumulative State - Query accumulated values across time windows, not just point-in-time snapshots

Visual Concept: Z-Axis Time Layering

• Recent attestations render on top, fully opaque
• Older attestations fade into background (z-axis depth)
• Click faded tiles to "jump back" to that time
• Rolling window follows scrubber position

AI Integration

• VidStream replay synchronized with attestation timeline
• AI inference results tied to both stream time and system time
• Graph evolution animated as video plays
• All three layers (video, AI predictions, graph) move together
• Future: semantic-aware temporal queries using ML-derived relatedness scores

Technical Foundation

• History depth managed by bounded storage
• Temporal aggregation across time windows
• Duration predicates enable accumulation
• Alternate timelines through attestation abstraction
• Rolling window filters for continuous monitoring

Future: Self-Describing Temporal Schemas

Vision: Ingesters attest their own temporal structure, making QNTX fully domain-agnostic.

Current (hardcoded):

# Code knows temporal field names
metadata: { "start_time": "2020-01-01", "duration_months": "36" }

Future (attested):

# Ingestor declares temporal schema via attestations
ingester:example -> has_temporal_field -> "start_time"
ingester:example -> has_duration_field -> "duration_months"
ingester:example -> temporal_unit -> "months"
ingester:example -> temporal_format -> "RFC3339"

# Queries discover structure dynamically
ax * over 5y since "2020"  # System reads attestations to know HOW to aggregate

Benefits:

• Zero hardcoded conventions - no assumptions about field names
• Cross-domain composability - any temporal data works without code changes
• Dynamic query planning - aggregation strategy derived from attestations
• Self-describing ingesters - like node_type attestations, but for temporal properties

This extends the attestation abstraction to time itself, completing QNTX's domain-agnostic vision.

Related Vision

• Continuous Intelligence - The paradigm that generates the history
• Glyphs - Attestable glyph state enables time-travel UI
• Fractal Workspace - Visualize time-travel through glyph manifestations

Roadmap

Phase 1: Temporal Aggregation ✅

• Duration accumulation across time windows
• Metadata-based temporal filtering
• Domain-agnostic query predicates (seconds, minutes, hours, months, years)

Phase 2: Semantic Awareness (planned)

• Weighted aggregation via relatedness scores
• Combined temporal + semantic filtering
• Multiple predicate AND logic

Phase 3: Temporal Overlap Detection (planned)

• Concurrent period merging to prevent double-counting
• Ongoing activity duration calculation