# Community Vision

A technical white paper that eats its own dogfood.

## The Idea

The StemeDB white paper makes claims. Dozens of them. Right now they look like facts—printed in authoritative academic prose. But they're claims: some well-sourced, some design decisions, some complexity assertions that haven't been benchmarked.

By highlighting claims in the paper itself, we demonstrate the core thesis visually. The reader sees: "Oh, this sentence isn't a fact—it's a claim, with a source (or without one), with a confidence level, potentially contested."

**The paper becomes a live demo of what StemeDB does.**

## What a Reader Sees

1. **Normal reading experience** — The paper reads like any academic paper
2. **Subtle highlights** — Claims have a dotted underline (amber, unobtrusive)
3. **Click to inspect** — Clicking a claim reveals:
   - Source attribution (with tier badges: T0 Regulatory, T1 Peer-reviewed, etc.)
   - Confidence level (0-100%)
   - Conflict status (Unanimous / Mostly agreed / Contested)
   - Notes on nuance or exceptions

## Why This Matters

Most whitepapers present claims as facts. This one presents claims as claims—with provenance. It's intellectually honest in a way that traditional documents can't be.

The reader learns what StemeDB is by experiencing it, not just reading about it.

## Technical Approach

1. **Phase 1: Static claims** — Claims are hardcoded in the React component with mock data (current state)
2. **Phase 2: API integration** — Claims are fetched from a running StemeDB instance
3. **Phase 3: Live updates** — As new sources are added or conflicts emerge, the paper updates

## Claims to Highlight

Priority claims to make interactive:

| Section | Claim | Why |
|---------|-------|-----|
| §1 | "Every mainstream database enforces single value per key" | Foundational claim, has nuance |
| §2 | "CRDTs assume replicas are authoritative copies" | Technical claim about related work |
| §4.2.1 | "RecencyLens is O(n)" | Complexity claim, needs benchmarking |
| §4.2.2 | "ConsensusLens groups by exact object value match" | Design decision with known limitation |
| §6.1 | "Append-only storage grows without bound" | Tradeoff claim |
| §7.1 | "Trust parameters are heuristics without theoretical foundation" | Honest limitation |

## Success Criteria

A reader finishes the paper understanding:
1. What StemeDB is (a claim-oriented database)
2. How it works (Lenses, source tiers, provenance)
3. Why it matters (disagreement is information)

...and they understood this by *using* it, not just reading about it.