My PGP fingerprint:
0xB589C6F74DCAE7A2E932A9773E0E253A69696969 Somebody, please quote. I will also archive this, and edit in a link. [Edit:
Archived.)
To show my love for Bitcoin, my primary key uses secp256k1. This was
added to GnuPG for the purpose of facilitating Bitcoin integration with gpg-agent. At some point, I plan to turn a few tricks with this feature.
PGP is sexy! ❤️ Bitcoin + PGP = orgasm. ;-)
I will not paste a copy of the key into this thread. The key itself may be obtained from
many keyservers, including
quite popular mainstays. The copies on keyservers can be updated with added/revoked subkeys/userids, new signatures, changed expiry times, etc. Those newly obtaining a key can always obtain the latest published version by such means; and those having an old copy can conveniently stay fresh with
gpg --refresh-keys.
There is no need to cargo-cult the copypasting and quoting of huge blocks of gibberish which fill huge threads with information which slowly goes obsolete.
The fingerprint is both necessary and sufficient to identify me with a key. Use the crypto, folks!
I will not sign a PGP message claiming ownership of this Bitcoin Forum account. Since this forum is important to me, I have already done better: My Bitcoin Forum uid is
declared as an OpenPGP userid on my key, secured by the appropriate self-sig. With this post, I estabish
bidirectional identity binding the best way it can be done here: Strongly, in the PGP → forum direction, and not-so-strongly in the forum → PGP direction.
Note, nullius has also added his forum userid to his PGP keys. You will need to
--refresh-keys those anyway, for updated expiry times on some subkeys.
PSA: Partly by way of
apologies to suchmoon, I made myself a custom
0x69696969 keyid with some hackery on my part, sipa’s secp256k1 keygrinder, and 1055.33375204 seconds on 1 core of the
CPU of an ancient laptop. IIUC, the
evil32 project used GPUs; but that’s not even necessary when using fast code for fast ECC algorithms.
I oft still see some people using 32-bit keyids, for purposes other than local management of keys already on their keyrings. Those still doing this,
please stop. Even a 64-bit “long” keyid is not sufficiently secure; the Bitcoin mining network now does more than 2
64 work every second! Only the full 160-bit fingerprint (a SHA-1 hash
0 of a specially constructed OpenPGP packet
1) should be considered reliable.
0. SHA-1 is considered generally “broken”, but remains resistant to preimage attacks. A full preimage attack would be needed to create a key with the same fingerprint as an already existing key. By contrast, it
may be feasible to newly create two different PGP keys with the same arbitrary fingerprint (
i.e. a collision); I am not sure. I also don’t see immediately anything that could accomplish for an attacker.
1.
RFC 4880 §12.2, as recently
explained to a snarky student.
