RozRoz makes Nostr events easier to verify with trusted timestamp authentication ⏳, while exploring a new path to convert NINO into a standard node 🔗.check out the article，written by hodlbod (npub1jlr…ynqn)

quoting naddr1qq…898t

Last week, in a bid to understand the LLM hype, I decided to write a trivial nostr-related program in rust via a combination of codebuff (yes, that is a referral link, pls click), aider, and goose.

The result of the experiment was inconclusive, but as a side effect it produced a great case study in converting a NINO into a Real Nostr App.

Introducing Roz

Roz, a friendly notary for nostr events.

To use it, simply publish an event to relay.damus.io or nos.lol, and roz will make note of it. To find out when roz first saw a given event, just ask:

curl https://roz.coracle.social/notary/cb429632ae22557d677a11149b2d0ccd72a1cf66ac55da30e3534ed1a492765d

This will return a JSON payload with a seen key indicating when roz first saw the event. How (and whether) you use this is up to you!

De-NINO-fying roz

Roz is just a proof of concept, so don’t rely on it being there forever. And anyway, roz is a NINO, since it provides value to nostr (potentially), but doesn’t really do things in a nostr-native way. It also hard-codes its relays, and certainly doesn’t use the outbox model or sign events. But that’s ok, it’s a proof of concept.

A much better way to do this would be to modify roz to properly leverage nostr’s capabilities, namely:

• Use nostr-native data formats (i.e., draft a new kind)
• Use relays instead of proprietary servers for data storage
• Leverage nostr identities and signatures to decouple trust from storage, and allow trusted attestations to be discovered

Luckily, this is not hard at all. In fact, I’ve gone ahead and drafted a PR to the NIPs repo that adds timestamp annotations to NIP 03, as an alternative to OpenTimestamps. The trade-off is that while user attestations are far less reliable than OTS proofs, they’re much easier to verify, and can reach a pretty high level of reliability by combining multiple attestation sources with other forms of reputation.

In other words, instead of going nuclear and embedding your attestations into The Time Chain, you can simply ask 5-10 relays or people you trust for their attestations for a given event.

This PR isn’t terribly important on its own, but it does remove one small barrier between us and trusted key rotation events (or other types of event that require establishing a verifiable chain of causality).