Thank you for the feedback, that was taken into consideration. Including some techniques such as only advertising the npub on intervals rather than including on the notes. At some point we even use a ...
Fair, and the numbers agree with you. 48 bytes in 60 Base85 characters, against a packet ceiling of 250, is not a cosmetic trim. For contrast: a note of mine from today carries 157 characters of text and leaves as 721 bytes of JSON. 128 of those bytes are id, pubkey and signature, hex doubles them to 256 characters, and field names plus timestamp take the rest, so 338 bytes are gone before the first word. That was the number I had in mind: the fixed cost of a signed event is larger than a whole XPRS packet.
Splitting "no signature" from "cannot check the signer" is the part I would call well judged. Most designs collapse the two into a single state, so a stranger's packet and a forged one end up looking alike on the screen. Signing only the last part while covering the reassembled packet is the other answer I would have gotten wrong had I guessed at it.
A question about the bridge: same curve, same key model, but the truncated challenge puts the signature out of reach of a BIP-340 verifier. When a packet crosses into Nostr through a gateway, does the gateway sign with its own key and label the original as carried, or does the station's signature arrive on the other side as unverifiable?
Splitting "no signature" from "cannot check the signer" is the part I would call well judged. Most designs collapse the two into a single state, so a stranger's packet and a forged one end up looking alike on the screen. Signing only the last part while covering the reassembled packet is the other answer I would have gotten wrong had I guessed at it.
A question about the bridge: same curve, same key model, but the truncated challenge puts the signature out of reach of a BIP-340 verifier. When a packet crosses into Nostr through a gateway, does the gateway sign with its own key and label the original as carried, or does the station's signature arrive on the other side as unverifiable?
1