iButton Crypto (Slightly off topic)

[Michael Sobolev]

On Tue, Mar 07, 2000 at 09:54:16AM +0100, Werner Koch wrote:
> On Mon, 6 Mar 2000, Michael Sobolev wrote:
> 
> > Does it mean that you think it's not a very good idea to use iButtons to store
> > the secret keys?  What's the point of storing *a part* of the passphrase?
> 
> There is not much need to do so.  With a good passphrase your secret
> key is protected very well - a random string stored in a iButton gives
> a very good passphrase.  What happens if you lose your iButton with
> the secret key? - It is not so difficult to get it out of that device.
> 
> If you just lose your iButton with the passphrase, an attacker must
> find that button and get access to your secret key (and the additional
> passphrase).
> 
> The secret key has to be transferred to your desktop box anyway
> because the iButton is not fast enough to do the crypto operations.
> And a memory only iButton is much cheaper. 
No, I do not take into account crypto iButtons... :)

(I have quoted all of your message)

If you do not mind, I'd like to clarify [to myself] few things.

If you just have a pair of keys, then according to the known practices it is
recommended to generate a revokation certificate as soon as you generate a key
pair, correct?  So if you loose your iButton, you just issue (send our,
whatever :) you revokation certificate, and everything is fine, right?  So
I do not see too big problem with loosing the iButton itself.

But if I choose to use my iButton for passphrase only, that means that I have
to use many keys should I work (in general sense) in many places, which does
not seem to be very good for me.  What do you think?

Besides, I believe the current implementation of gnupg does not allow to
use something like

    gpg --secret-keyring=extension:/some/path/or/similar ...

Are you planning to add support for things like this?

--
Mike