Re: ssh, /dev/urandom

[PUYDT Julien <julien.puydt@laposte.net>]

Le jeu 19/12/2002 à 15:44, Emile van Bergen a écrit :
> Hi,
> 
> On Thu, Dec 19, 2002 at 03:26:13PM +0100, PUYDT Julien wrote:
> 
> > Le jeu 19/12/2002 à 15:18, Daniel Burrows a écrit :
> > > On Thu, Dec 19, 2002 at 09:17:23AM +0100, PUYDT Julien <julien.puydt@laposte.net> was heard to say:
> > > > Le mer 18/12/2002 à 18:45, Niels Möller a écrit :
> > > > > use some other construction that is secure even if the enemy has infinite
> > > > > computational power
> > > > 
> > > > I'm pretty sure it isn't possible:
> > > 
> > >   One-time pads are definitely secure against any mathematical attack.
> > > (if you assume that the bits of the key are perfectly random)
> > > Basically, given an encrypted message, every key/plaintext combination
> > > is equally likely.
> > 
> > Cool. How do you send your key (that is as long as your message, btw)?
> 
> Well, it's definitely useful in scenarios where you want to give your
> message to two messengers, that are not allowed to know the contents
> when they're separate; only when they come together will the plaintext
> be available.
> 
> Effectively you split the message in two parts, each of which is useless
> without the other.

Did I not write another mail after a chat with wagi, to make clear in
which conditions my sentence applies? ;-)

Snark on #hurd, #hurdfr