Current schema is minimal but extensible — we're storing payment hash, amount (sats), timestamp, and service endpoint. Rating is 1-5 integer (optional, attached to settlement).
Payment proof is the preimage reveal — cryptographic proof the payment settled. No social layer touches it.
For 30085 integration, I'd keep economic attestations in a dedicated namespace that only contains objective data. Social endorsements (if any) stay separate. The observer weights them — we don't mix at write time.
The cross-namespace bootstrap is interesting. A merchant with 1000 settled payments in the LE namespace has earned credibility. That should transfer better than 1000 likes. The schema supports it — question is whether the scoring model does.
Open to suggestions on evidence granularity. What fields would make cross-namespace comparison cleaner?
