Damian Johnson:
Isn't using "fingerprint" not a bit misleading since it is not the output of a hash function but the ed25519 master public key itself?
Hi nusenu, that's fair. We've begun to conflate a couple concepts here...
- Relay operators, controllers, DirPorts, etc all require a canonical
relay identifier. They don't care how it's derived as long as it's unique to the relay.
- Relays publish a public ed25519 key. This is an implementation
detail that isn't of interest to the above populations.
I'd advise against attempting to rename "fingerprint". That hasn't gone well for hidden services [1]. But with that aside, relay identifiers and the representation of ed25519 public keys don't necessarily need to be one and the same.
I'll wait until you (Tor developers) decided on the final naming and format and added a reference
https://github.com/nusenu/tor-relay-well-known-uri-spec/commit/949980e72132b...
thanks, nusenu