non-standard key types

Atom 'Smasher' atom-gpg at suspicious.org
Mon Nov 3 00:12:34 CET 2003


ok, the default setting for generating a new key-pair is:
   (1) DSA and ElGamal (default)

re-reading the DSA section in applied cryptography, i was wondering about
using a master key that can be made bigger than DSA's limit of 1K.

then of course, i started playing (to see if it could be done without a
fight) with both RSA and ElGamal sub-keys (for encryption).

here's the keyring that i generated:

pub  2048R/D8E2A23F 2003-11-03 testing (testing RSA master key) <x at x.xxx>
sub  4096g/86EE3F36 2003-11-03
sub  4096R/83F2C8E8 2003-11-03

note: on freeBSD 4.8, gpg 1.2.3, 1.2G Duron, with `rndcontrol` well
seeded, the keys took less than 2 minutes each to cook.

anyway, i have 2 questions:
1) would a 2K (or larger) RSA master-key cost me anything in terms of
	interoperability?
2) given the current state of crypto research, is it considered
	more secure to use an RSA or an ElGamal sub-key (assuming
	they're both the same size)?

before someone asks me what color my tin-foil hat is, i fully understand
that this is far beyond practical applications, and of interest only for
either academic or clinically paranoid ends... i like to think i fall into
the former category  ;)

any other thoughts on ~not~ using the defaults for generating new keys?


 	...atom

_______________________________________________
PGP key - http://smasher.suspicious.org/pgp.txt
3EBE 2810 30AE 601D 54B2 4A90 9C28 0BBF 3D7D 41E3
-------------------------------------------------

	"The more laws, the less justice."
		-- Marcus Tullius Cicero De Officiis




More information about the Gnupg-users mailing list