GPG testing...

bscott at ntisys.com bscott at ntisys.com
Sun Dec 29 22:45:10 EST 2002


On Sun, 29 Dec 2002, at 10:24pm, gnhlug at sophic.org wrote:
>> That is rather missing the point.  The reason non-repudiation is desired
>> is that it means one cannot say, "I never sent that."
> 
> No, that's only one reason.  The other reason is to say, "we can prove
> that you sent this."

  No, that's authentication.

  Non-repudiation, by definition, means one cannot repudiate any given
message.  Repudiation, by definition, is the act of denying you sent a
message.  If signing is voluntary, you can always assert that you simply did
not sign a given message.  Thus, voluntary signing does not provide
non-repudiation.

> In such a case, you definitely want them to be digitally signed, and
> reject anything that isn't.

  If anything which is not signed is rejected, then signing is not
voluntary, is it?

  Case in point: This discussion originated as a discussion about using
digital signatures to counter spam.  Since digital signatures, on today's
Internet, are relatively uncommon, they do not provide non-repudiation.  
Thus, digital signatures cannot be used to prove one did not send a given
spam.

  Now, I am sure someone will say, "If you sign all your messages, then the
unsigned spam will be suspect, because it lacks your digital signature."

  That again misses the most fundamental aspect of security: Security is
entirely about trust.  Someone sending illegitimate mail is, almost by
definition, not to be trusted.  Thus, if you are suspected of sending an
illegitimate message, the fact that you nominally sign all your messages
does not impart trust.  Indeed, one who regularly traffics in illegitimate
messages would be rather more likely to sign all their legitimate mail.  
Meanwhile, if you can, by other means, prove you are trustworthy, the
digital signature becomes superfluous.  We already know you are trustworthy;  
thus, we don't need a digital signature to know you did not send the
illegitimate message.

-- 
Ben Scott <bscott at ntisys.com>
| The opinions expressed in this message are those of the author and do not |
| necessarily represent the views or policy of any other person, entity or  |
| organization.  All information is provided without warranty of any kind.  |




More information about the gnhlug-discuss mailing list