-
@ hoppe
2025-04-06 03:03:56
Thank you for the interesting article. It was very helpful in understanding the nostr ecosystem.
I have one question to confirm if my understanding is correct.
Assuming all clients adhere to the Outbox Model (NIP-65), and all relays store and serve requested notes completely, users should theoretically be able to find 100% of the posts authored by people they follow and the reactions to those specific posts without guessing or manually adding relays (ignoring the bootstrapping problem of how to initially find the NIP-65 event for someone I follow, let's assume that's solvable through reasonable guessing or indexers).
However, my understanding is that finding an original post (let's call it Post A) when someone I follow replies to or quotes it cannot be achieved 100% solely through the Outbox Model of the person I follow. It seems dependent on NIP-10 hints (p and e tags).
Specifically, is this process correct?
When my follow replies/quotes, their event must include hints (like a p-tag for the original author, let's call them Author B, and an e-tag for Post A).
My client finds the reply/quote on my follow's write relays (as per NIP-65).
My client extracts Author B's pubkey (from the p-tag) and Post A's event ID (from the e-tag).
My client then needs to find Author B's NIP-65 event (facing the bootstrapping issue again).
Then, my client must query Author B's write relays to fetch Post A itself (using its event ID).
And additionally, query Author B's read relays to find other reactions (likes, zaps, other replies) to Post A (since Author B is tagged in those reactions).
Is this multi-step process, relying on hints and the original author's NIP-65 relays, the correct way this interaction works within the Outbox Model's scope?