Foraging Platform Design
Core Principles
- No Following, No Global Graph: Users cannot follow each other. Influence is room-local and decays with inactivity.
- Rooms Never Truly Die: Communities go dormant but are not deleted, preserving historical footprints for discovery and revival.
- Eviction Is Hygiene, Not Punishment: Inactive members are cycled out after long decay periods, freeing slots while signaling value through scarcity.
- Scarcity Drives Intimacy: Rooms have hard caps (with narrow ranges, e.g. 50–100 members). Large, thousand-member rooms are impossible.
- Discovery Feels Like Foraging: Users browse activity charts, not infinite feeds. Joining a room feels like uncovering a living ecosystem, not subscribing to a content firehose.
Membership Lifecycle
- Join → Active → Dormant → Evicted → Cooldown → Rejoin
- Eviction is automatic after inactivity passes thresholds.
- Cooldowns prevent slot-squatting and bot churn.
- Away tokens (limited per year) allow pausing decay during vacations or emergencies.
Room-Level Policies
- Hard Caps: Size fixed within platform-mandated range.
- Participation Required: Seats are not permanent; all members subject to decay.
- No Selling/Clout Farming: Rooms are not promotion channels; feedback system flags low-value or overly promotional dynamics.
Discovery & Search
Discovery uses a bubble-style foraging UI, with optional charts/tabs for power users:
Bubble Canvas (default)
- Bubbles = rooms. Size encodes current activity; border thickness encodes capacity pressure; color hue encodes mode (voice/live-text/async), and saturation encodes freshness.
- Physics: bubbles drift toward your weighted preferences (freshness, intimacy, depth); dormant rooms sink but remain discoverable.
- Hover reveals: live headcount, queue length, 7/30d sparkline, “vibe” tags (e.g., balanced, mentor-heavy).
- Click opens the room panel with tabs (below).
Tabs in Room Panel
- Activity: historical and real-time participation chart (like trading candles or a sparkline deck).
- Conversations: recent sessions with avg participants, duration, mode breakdown.
- Footprint: dormancy history, revival streaks, talk-time equity metrics.
- Queue/Capacity: current cap, seats free, estimated queue ETA band.
- Feedback: randomized, LLM-assisted impressions aggregated and de-gamed (never user-initiated).
- Similar: nearby bubbles by topic/mode/size.
Additional Discovery Signals
- Conversation Metrics: avg participants per session; voice vs. text ratio.
- Dormancy Status: active → cooling → dormant, with revival badges.
- Room Footprints: public activity records that never delete.
- Balanced Surfacing: intentional mix of dormant legacy rooms and new sprouts.
Note: “bubble” here is a UI metaphor for surfacing/float dynamics — not the bubble‑sort algorithm.
Feedback & Reputation
- Contextual Reputation Only: Scores decay and are room-specific.
- Feedback System: Users provide randomized, low-effort evaluations that inform room footprints without gaming.
- No Global Rankings: Discovery is based on browsing, filters, and user-set weights, not universal leaderboards.
Outcome
- Intimacy Over Scale: Communities remain small enough to form real bonds.
- Scarcity Makes Participation Meaningful: Seats are valued because they are limited and revocable.
- Serendipity Returns: Users discover rooms like explorers, not consumers.
- Evolutionary Dynamics: If excluded, users can fork and start their own rooms, growing the ecosystem.
Related Reading
- PK-ToDo Add link Bubble-Style Discovery UI: a visual foraging interface where rooms appear as interactive bubbles that surface based on your weights (freshness, size, mode), not the bubble‑sort algorithm.