[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Subject signing redux (was: Re: Mary is Mary)
On Wed, 25 Jun 1997, David P. Kemp wrote:
-> I hate to beat a dead horse, but this again comes down to the premise
-> that the verifier is central. It's given lip service, but seems to be
-> forgotten every so often.
-> Mary = subject.
-> Jon = issuer.
-> Ed = verifier.
-> Ed wants to know if Mary is a lawyer.
-> Jon says Mary is a lawyer.
-> Ed either trusts Jon or he doesn't. If he trusts Jon, and Jon lied,
-> then Ed has the basis for a lawsuit against Jon, not against Mary.
-> Mary doesn't have to countersign anything.
This point is clear, but it is not the issue. The attack is also
1. Engineer, don't suppose:
As once said, law is no substitute for engineering. Ed should not be
jeopardized by a system that lets Jon create legal trouble -- when it's an
engineering task to avoid it ... just include Mary's signature. I would
like to hear what's so bad about Mary signing -- protocol-wise --
that would make it so difficult to close this security hole..
2. The attack:
The attack is "framing Mary".
If Jon lied how can Mary prove she did NOT agree and now is just
trying to avoid trouble? How can Mary prove she did not lie? If she would
be required to co-sign, then she could *prove* she did not lie.
Further, Jon can present a falsified statement from Mary, where she
supposedly sends him a note saying she has just been authorized by ABA.
Since Jon trusts Mary, he doesn't check that (i.e., in the same way that
Ed would not call up Mary and ask her if Jon's cert is correct) -- he
would say in court.
So, Mary had her name used (possible constitutional rights?) without her
authorization, to commit her to an obligation -- and she can't prove
3. the nuisance:
Why should Mary have to defend herself -- and spend time and money on it?
4. Another example:
It's in the SPKI draft itself!
Dr.rer.nat. E. Gerck email@example.com
P.O.Box 1201, CEP13001-970, Campinas-SP, Brazil - Fax: +55-19-2429533