From: eric@remailer.net (Eric Hughes)
Date: Wed, 8 Feb 95 20:10:38 PST
To: cypherpunks@toad.com
Subject: Re: Hastur CT status
In-Reply-To: <199502082009.TAA22774@seabsd.hks.net>
Message-ID: <199502090408.UAA24470@largo.remailer.net>
MIME-Version: 1.0
Content-Type: text/plain


   From: A Loose Affiliation of Millionaires and Billionaires and Babies <cactus@hks.net>

   For instance, I've asserted several times here that X.509 keys
   can be fully modelled as special cases of PGP web-of-trust keys
   with one additional field, the expiration time.  Nobody has flamed
   me, but nobody has agreed with me... since I've only read the X.509
   spec and never actually used them, I'd like some assurance that I'm
   not missing some subtleties in this approach.
   
The only real question about a particular string of bits claiming to
be a public key of a certain persona is whether the operator trusts
that the key does belong to that persona.  PGP and X.509 models both
provide their own kinds of assurances to individuals who might use a
key.  The relation between the user and the claim of ownership is the
important relationship.  Any sort of key certificate, of whatever
sort, is merely an aid to gaining trust.

Key certificates don't prove ownership.  Key certificates transfer the
need for trust in the key to the need for trust in the certificate.
Put another way, a key distribution system allows a user to trust
something harder to fake than a single key.  The transfer is the
critical point here; instead of trusting one small thing, you can
trust one larger thing.  We hope that the larger system is worthy of
our trust.

Neither PGP nor PEM is a general purpose key distribution system,
although PGP is more general than PEM.  Both have their various
arbitrary and capricious policies hardcoded into both spec and source.

I would recommend, Todd, that you not try to unify the various key
distribution systems.  It's premature.  Rather, provide a local policy
hook for the user (and this is _not_ just the sysadmin, as you know)
to specify how much trust pertains to each given keydist system, and
of what idiosyncratic sorts.

Eric