Thread

I still think that one Identity key and multiple device specific keys isn't that complicated. It's what we have now, except you validate in-event identity claims, and when you see a new revocation event prune the disavowed device events.

GitHub

NIP-102: Subkey Attestation by ynniv · Pull Request #1450 · nostr-protocol/nips

This NIP defines a way to separate identity from authentication using hierarchical deterministic (HD) keys. This allows people to use one key pair ...

Sounds interesting, look forward to testing it out when its a bit more built out. One of the strengths of Nostr is simplicity, a user signed 5 field object is all you need to know. And if you want to get fancy, a websocket subscription and some event filtering.

> sig-chains - you know if you're missing content in PZP You could do the same on nostr if anybody actually cared about this. Just include a "prev" tag with the ID of the previous message in the chain you're building. You could also have a "last" tag or something to indicate the chain is complete. Clients could render it as "message 3/9" or whatever, like people do manually on Twitter. In fact, PZP's replication appears to *rely* on these hash chains, since it uses hash graph replication:

Codeberg.org

pzp-sync

PZP replication using Kleppmann's hash graph sync

I did some analysis of this sync method and if you tried to use it for non-linked data you can easily cause it to degrade into a worst-case behaviour:

GitHub

GitHub - hoytech/automerge-poison

Contribute to hoytech/automerge-poison development by creating an account on GitHub.

IMO RBSR/negentropy is better, one of the reasons being you can sync arbitrary collections of (unlinked) data, for example all kind 0 notes, or all notes with a certain hash tag.