# Ethcc proposal

Track: `5. Cypherpunk & Privacy
Good luck spying on this..`

## Building on Private State: Read & Write Privacy in practice

### Short summary (≤250 chars)
ZK hides state, but production apps still have to read and write it. Using real chains as case studies, we compare real patterns for scanning, proving and delegation, and their UX/privacy costs.

### Full description (≤1000 chars)
On public chains, indexers see everything. On private systems, you still need to find and update encrypted state without leaking who you are, what you hold, or which dapp you care about.

This talk looks at how existing systems actually do it, from a developer’s point of view.

**Pattern 1: Ethereum privacy overlays (Railgun, Tornado/Privacy Pools, Hinkal)**
These keep Ethereum’s state model and add privacy on top. Railgun holds balances in zk-SNARK pools with relayers submitting transactions; Tornado and Privacy Pools use mixer-style anonymity sets; Hinkal uses stealth-like addresses for private DeFi. Integration is straightforward for web and mobile wallets, but you inherit Ethereum’s indexing, mempool behaviour and endpoint trust assumptions.

Pattern 2: Local proving + distributed tx provers (Shield)
You can prove locally on your device for maximum privacy, or delegate to a prover (run by 3rd party) for better UX. User chooses between trust/performance.
For indexing, users can allow 3rd party indexer to see their private transactions instead or just local scan.

~~Pattern 3: Client-side proving with local state (Aztec)
Private state lives as encrypted notes locally. You maintain a local database of notes and prove on-device. Privacy-by-default, but high client CPU and latency. No delegated proving option.~~

We map each pattern's costs: compute requirements, latency, fees, and what observers learn. Indexing strategy, proving model, and how state ownership differ across chains.

### Impact (≤500 chars)
Attendees get a reusable pattern library for private state—tied to concrete implementations (Zcash, Aleo, Aztec) but applicable anywhere. They'll understand which indexing and proving models fit their threat model, what they trade in UX and privacy, and leave with understanding of "private" chains before betting their product on it.