David Picon Alvarez
Fri Feb 21 11:00:01 2003
-----BEGIN PGP SIGNED MESSAGE-----
> The big problem with key sharing algorithm is: where do you do the
> computation? Where do you assemble the key, and how do you guarantee
> that the owner of that computer does not keep a copy of the assembled
> key? IIRC, the classical key sharing algorithms work by distributing a
> secret (without any knowledge about what that secret is), the assembled
> secret is then used as key.
Hmm. No, that's not quite what I'd like to do. If the secret gets computed
somewhere then you could as well distribute pieces of a passphrase or
> I'm told this problem can be solved, there are
> algorithms where you can
> distribute the computation without the whole key ever being assembled.
> But I'm not into that, so you'll have to do your own research
This is more like it. I've been browsing around Google, but with not very
conclusive results. I guess I'll have to keep on going.
> require one designated computer to be trusted, you've lost.
I agree 100% with you.
> Is that still your insanely big key? I dare not verify your signatures,
> because gpg is sooooo slooooooowwww. Really defeats the purpose of
> digital signatures...
Hmm. It's not insanely big, it's 4096, but I guess you say it because of the
ElGamal sign and encrypt thing. I don't know, I don't want trouble with
subkeys, and especially I don't want DSA because of the 1024 bit limit.
Maybe I'm being too stubborn about this and I should go to RSA 4096/4096.
However, saying it defeats the purpose of digital signatures is quite an
-----BEGIN PGP SIGNATURE-----
Comment: This message is digitally signed and can be verified for authenticity.
-----END PGP SIGNATURE-----