Re: Certificate validity issues

[lmccarth@cs.umass.edu]

some hairboy ;) writes:
> I don't see any that there could be any sort of automatic way to 
> distinguish between the necessity for recipient-signed vs. 
> recipient-unsigned certificates; One would assume that recipient-unsigned 
> certificates would be effective only when issued from a high-reputation 
> source.
> 
> Some method should be required to allow user decisions as to this, but 
> whatever system is designed should definitely allow for both types, and 
> *require* relevant software to handle both types.

I think this is a policy issue, not a technology issue. In general, key
owners should be able to sign certificates that say arbitrary things, 
regardless of whether or not some third party agrees that the statement
being signed is "true". Unless you believe in libel, slander, etc., which
raises policy issues of a different sort entirely.

The John Birch Society can sign something claiming that I'm a member. 
A policy that says that's sufficient information to conclude that I *am* a 
John Bircher is a broken policy. A policy that requires both the above and
a signature by me on the same certificate, among other things, is viable.

> There's actually a third sort of important certificate the system should 
> handle:
> 
> (3) I might be a member of a secret society; I might need a membership 
>     certificate to get access, say, to certain web sites. The system 
>     should allow a "secret certificate," readable only by the issuer.

Your description is a bit ambiguous. Do you really mean "readable", or
do you mean "verifiable" ?

If the former, it sounds like this could be handled by distributing 
signed certificates encrypted with the issuer's public key. 

If the latter, you could distribute signed certificates but not the public key
needed to verify the signatures on the certs.

It might be handy to use a hybrid approach, such as distributing signed
certificates that refer to encrypted entity names (suitably padded with
random bytes to avert trial encryption of plaintext guesses).  

-Lewis	"You're always disappointed, nothing seems to keep you high -- drive 
	your bargains, push your papers, win your medals, fuck your strangers;
	don't it leave you on the empty side ?"  (Joni Mitchell, 1972)