# Boreal launch plan v0.1

User decisions, 8 November 2026: public testnet first with free test coins;
website purchases should eventually support both cards and cryptocurrency.
"Better than every coin" is an ambition, not an established claim. Publish
measurable improvements and failures against explicit baselines instead.

## Implemented staging milestone

A separate pinned-genesis research chain, versioned ML-DSA-65 transaction domain
and address prefix, bounded loopback transport, independent replay/fork selection,
atomic durable node state and three systemd-managed node processes on one host.
Public HTTPS snapshots/source permit independent verification. This is a read-only
public preview, not open peering, a faucet or a production network. Research
SHA-256 difficulty is intentionally trivial and separate from candidate GPU PoW.
Default rewards are unspendable and cannot fund sales or a faucet.

## Next concrete milestones

1. Deliver the previously requested CUDA job adapter and reproducible build path,
   then verify integer-exact solutions with an independent CPU implementation.
   Record actual hardware/toolchain and bounded runtime; no invented B300 results.
2. Build official iOS/Android wallets and Windows/Ubuntu desktop wallets with
   full validating nodes. Separate node networking from signing. Verify a reviewed
   key-protection/authorization design and CUDA-mined private development rewards
   arriving at a private wallet public address. Native apps remain unbuilt.
   Biometric modality remains undecided; see MOBILE-WALLET-SECURITY-v0.1.md.
   Do not move private keys or biometric templates to the server.
3. Harden the research network for public bidirectional participation: versioned
   wire protocol, domain/genesis handshake, bounded inbound messages/connections,
   admission and rate limits, TLS deployment/reachability, mempool reservations,
   transaction/block relay, partition/reorg/stale-job recovery and long-lived
   bounded storage. Keep existing chain and testnet-v1 identity protected.
4. Reproduce node synchronization on independent hosts/implementations. Produce
   signed reproducible client releases, a test-only wallet/faucet and explorer.
   Publish testnet reset policy, capacities and remaining attack assumptions.
5. Research a dependent shared-epoch PoW candidate with recomputation and verifier
   budgets. Candidate v2 was rejected: independently addressable cells permit a
   cheap storage shortcut. No final protocol, ASIC-resistance or GPU advantage claim.

## Bridges

BOREAL on Ethereum and Solana means escrow-backed representations sharing one
aggregate backing invariant. Mint/burn accounting exists only in a local model.
Before test deployments: canonical chain/domain/address encodings, actual escrow
ownership, finality verification, route-bound replay protection, integer atoms,
paused/reorg/incident behavior and independent adversarial review. A mint authority
or operator assertion is not a source-chain proof. No contract or mint exists yet.

Bitcoin is BTC moving into/out of Boreal, using separately backed satoshi units
and BTC redemption. First specify a separately bounded multi-asset ledger and
custody/proof model. Do not reuse BOREAL escrow as BTC backing, conflate a BTC sale
payment with a BTC bridge, or market a federation as a trustless bridge. Tests
must cover duplicate outpoints, reorgs, reserve reconciliation, pending redemption
and recovery. The custody/proof trust choice remains unresolved.

## Website purchases: both payment rails, unavailable until configured

Collect sale asset/network, inventory source, integer price/quote-expiry rules,
seller identity, operating and supported buyer jurisdictions, provider account,
refund policy and delivery model before enabling checkout. No collecting funds
while these decisions are missing. No mainnet coin exists to deliver today.

Card flow: provider-hosted payment UI; no card details through WordPress. Validate
webhook signatures, amount/currency/order binding and event replay protection.
A browser success redirect never authorizes coin delivery. Design reconciliation,
chargebacks, refunds, retries and reserved inventory before accepting payment.

Crypto flow: exact asset/network-specific invoices, integer smallest-unit amounts,
unique order binding, expiry, reorg/finality policy, under/over/late payment handling
and explicit refund validation. A pasted transaction hash is not proof of payment.
Never expose seed phrases, signing keys or an unrestricted mint/delivery key to
WordPress or browser code. BTC checkout does not implement the BTC bridge.

Both flows require a durable order state machine with idempotent fulfillment,
separate inventory reservations, delivery transaction reconciliation and failure
recovery. Establish merchant eligibility and sale obligations for the actual seller
and jurisdictions with the selected provider and qualified advice. No legal
compliance, price appreciation, launch date or security guarantees are claimed.

## Public claims

No monetary value, token sale, functioning bridge, public faucet, working final
GPU PoW or audited production wallet exists. Public pages should show verified
milestones, accurate data age and exact remaining gates. Never invent adoption,
security rankings, pool partnerships, NVIDIA endorsement or external social posts.