The protocol doesn't support multiple signatures. That would create an invalid event.
Here is a howto for the musig2 scheme https://github.com/fiatjaf/nak/blob/a6509909d0d83405a5be57fbc1831f6bb96b316e/README.md?plain=1#L91
Discussion
Thanks - I was thinking more just like a 'coauthors' tag that was the list of (npub, sig) tuples beyond the first author
The tags are part of the entropy of the signature, so you can't include a signature in the tags without changing the signature, ad infinitum
ah of course
That is what we have, as you can have n author or p tags, but that isn't the same thing as having multiple people sign the same event with their nsec. That would be cryptographic and break protocol.
It isn't clear which solution you prefer.
I wanted something cryptographic that would be like when you submit a paper for publication. Ideally, each author is listed separately, but they have a signature proving they weren't added to the paper (note) without their consent. I suppose though that all the coauthors could just do a multisig at that time they finalize the paper for submission and send it off?
Multisig adds a lot of complication.
We're doing peer review and co-authorship with approval/attestation events. Sort of the way you can recommend someone else's 30818 wiki page.
One npub writes it and lists all of the authors, publishes, and the authors sign-off on that, and the client gives their entry in the author list a verification ✔️
Spec? :eyes:
When it goes live, for nostr:npub1l40cx7mc8cr23qamy6xhjx4gng9wmuczj2d8ertvw98w0aak9x3qyxmh7n .
It's a major feature, so we need to be clever about it.
1988 would be a good number for it, though, as it's akin to a label or report. Sort of the inverse of a report.
Something simple, as it's just document signing. The pubkey just needs to match one of the document's p-tags. Could use it for anything, really.
{
"kind": 1988,
"tags": [
[ "role", "author" ]
]
}
Sorry, forgot the e tag. No a tags, obviously, as you want to sign-off on a particular version of a publicatio.
Yup, was about to say 🎯
Would somehow be nice to match the Relationship Status spec for consistency.
With:
- The one letter tags
- The ability to sign of on multiple roles with one event
And that Relationship Status nip needs:
- a better name (profile labels or sth)
- ability to also include a :badge: Badge instead of the label string (n)
Ok cool.
Do you leave the "role" open? Or is being an "author" the only option?
How do you send out the "please sign this for confirmation" thing? Via Nostr Mail?
There are different roles, like author, reviewer, approver. You could sign with multiple roles. Role depends upon the review stage.
It's a bit complicated, as it needs to match university and industry standard, but the basic idea of a 1988 sign-off is more universally applicable.
Well, you get notified when an event includes you in a p-tag. Don't need a second notice.
Sweet!
I think that last part might be naive. But we'll see, depends on the apps that back the interaction up.
