This interface evolution note has been archived. The original document is available in the archive folder.

Archived copy: archive_docs/INTERFACE_EVOLUTION.md

For the active interaction model and UI decision summary, see GRAPH_INTERFACE.md and README.md.

Feature	Status	Notes
Graph-based spatial browsing	✅ MVP	Force-directed nodes + edges, force physics
Map view as primary interface	✅ MVP	Graph canvas opens on app launch
Tabs as connection list	✅ MVP	Tabs in detail window show edges to/from node, ordered chronologically
Multi-window support	✅ MVP	Windows 11 snap-like layouts, side-by-side browsing
Personal history view	✅ Phase 2	Can reconstruct graph from browser history, personal OPTE-like view
Search/filtering	✅ MVP	Live search, filter by title/URL/tags
Clustering & grouping	✅ Phase 2	Detect connected components, collapse/expand

Feature	Old Plan	Current Plan	Timeline
Blockchain wayback machine	Full feature	Research + prototype	Phase 3+
Sharing active tabs/webs	P2P sync	Simple P2P + event-based merge	Phase 3 research
Live collaborative browsing	Sandbox instance of others' webs	Future network layer	Phase 4+
OPTE-style visualization	3D/5D archive overlay	Simpler: 3D toggle (Phase 2)	Phase 2–3
Time as scrollable dimension	5D (x, y, z, time, validation)	Time as edge metadata only	Phase 2+
Consensus validation	Two-tier (blocks + relationships)	Research: event-based validation	Phase 3
Domain-level abstraction	Primary grouping unit	Implicit (clustering by domain)	Phase 2–3
IP/location positioning	User localization (where is "here"?)	Not in current plan	Future

---

The older notes sketched an ambitious vision:

1. Personal history as a personal OPTE (Internet Map-like visualization)
2. Shared webs — Access other people's active tab collections as sandboxed instances
3. Blockchain validation — Prior compiled versions of sites via wayback, validated by consensus
4. Multi-dimensional — Space (x, y, z) + time (scroll history) + validation strength
5. Consensus organization — Sites validated and organized by community consensus, not just traffic
6. Domain-centric — Group sites by domain (simpler relationships than all-pairs)

Phase 1 (Weeks 1–8): Single-user force-directed graph

• Create/navigate nodes (URLs)
• Edges are simple connections (chronological, user-created)
• Detail windows show adjacent nodes as tabs
• Multi-window snapping for side-by-side exploration
• Save/load graph as JSON

Phase 2 (Weeks 9–16): Personal features

• Persist graph to disk + bookmarks import
• Clustering (auto-detect groups)
• Filtering (by domain, date, tags)
• 3D visualization (optional)
• DOM extraction (capture page snippets)
• Export (JSON, HTML, PNG)

Phase 3 (Weeks 17–24): Network & ecosystem research

• P2P sync model (event-based: node added/removed/updated)
• Shared graph merge (simple decentralized sync)
• Tokenization research (user data ownership, compensation)
• Validation model (how to trust shared graphs?)

---

Old: Access prior validated versions of sites compiled from a blockchain Now: Phase 3+ research; MVP focuses on current-state browsing

Rationale:

• Blockchain adds complexity (validation, chain management, consensus)
• Current version capture (DOM extraction in Phase 2) covers snapshot use case
• Full wayback requires upstream validation infrastructure

Old: Click a person's shared web → live sandbox of their browsing Now: Phase 3 P2P sync → can share graph structure + validate edges

Rationale:

• Sandbox instance requires container tech + ongoing sync
• MVP: Personal graphs only (no sharing)
• Phase 3: Simple event-based sync (graph structure, edge timestamps)
• Future: Layer in validation/sandbox after P2P stabilizes

Old: 5D visualization (x, y, z, time, validation); scroll through history Now: Time stored as created_at on edges; 3D optional in Phase 2

Rationale:

• 5D is hard to visualize and implement
• Timestamp on edges enables chronological tab ordering (MVP requirement)
• 3D (Phase 2) provides optional depth; time through history becomes "navigate back in graph"
• Can revisit true time-dimension visualization in Phase 4+

Old: Two-tier validation (blocks + site relationships); consensus from community Now: Phase 3 research; MVP has no validation (single-user)

Rationale:

• Consensus requires network + incentives
• MVP is local-only; can't validate without peers
• Phase 3 explores event-based validation (edge strength = how many users saw connection)
• Avoid premature blockchain commitment; MVP proves graph model first

Old: Primary grouping: domains as zones; domains relate to other domains Now: Nodes are URLs; Phase 2 auto-detects clusters by connectivity

Rationale:

• Site-level granularity more flexible (can mix domains in same cluster)
• Clustering algorithm (connected components) provides domain-like grouping automatically
• Can surface domain grouping in Phase 2 if useful

Old: "Where is 'here' online?" — IP-based, IRL location, site connections Now: No positioning; graph position is physics-driven, not location-based

Rationale:

• User location (IP, physical) adds complexity without clear UX benefit for MVP
• Graph layout driven by relationships, not geography
• Could revisit for Phase 3+ (e.g., "nearby" peers in P2P network)

---

Complexity	MVP Avoids	Phase 3+ Adds
Consensus	None; single user	Event-based validation + stakes
Sync	No sharing	P2P merge of graphs
Archive	Current snapshots	Blockchain wayback?
Visualization	2D + optional 3D	5D (time scrolling), OPTE overlay
Geography	None	IP/location positioning
Sandboxing	N/A	Live collaborative web instances

---

• Graph is primary interface ✅ (unchanged)
• Spatial/force-directed layout ✅ (core MVP)
• Tabs reflect graph structure ✅ (connection list model)
• Personal knowledge management ✅ (history import, clustering)
• Eventual ecosystem play ✅ (Phase 3+ P2P + tokenization)

• Blockchain validator → Local timestamp-based edge validation
• Shared sandbox webs → Simple P2P graph sync
• 5D visualization → 3D toggle + time metadata
• Consensus mechanism → Event-based (revisit in Phase 3)
• Domain zones → Implicit clustering algorithm

---

After MVP is solid (Phase 1–2), investigate:

1. Validation model: How do we trust edges in a shared graph?

   • Options: Majority vote, stake-weighted, timestamp-based, user reputation
   • Blockchain vs. decentralized append-only log vs. event sourcing

2. Sandbox & live sharing: How to give access to someone's live web?

   • Options: WebContainer API, Docker, virtual browser, HTML archive + snapshots

3. Wayback integration: Can we leverage existing wayback machines + blockchain?

   • OPTE-like (visual archive), but simpler
   • Or build blockchain archive incrementally as users visit sites?

4. Tokenization: How do we compensate contributors/validators?

   • Data marketplaces, governance tokens, reputation systems
   • Avoid speculation; focus on legitimate user incentives

5. Time dimension: Revisit after Phase 3 P2P is working

   • Could allow "history scrubbing" (view graph at point in time)
   • Or explicit version control per graph snapshot

---

Start with Phase 1 MVP exactly as documented — it's sound architecture that doesn't preclude the bigger vision. The deferred features (blockchain, sharing, consensus) are additive layers that can be built on top of a solid single-user graph foundation.

Avoid premature implementation of:

• Blockchain validation (wait for clear use case in Phase 3)
• Sandbox instances (wait for P2P + user demand)
• 5D visualization (3D + metadata is sufficient for MVP)

The MVP proves the core insight: force-directed graphs work for spatial browsing. Once that's solid, the network + ecosystem features become tractable engineering problems rather than architectural risks.