On Mon, 1 Jul 2002 10:45:44 -0400 (EDT) Richard Welty <[EMAIL PROTECTED]> wrote:
RW> the complexity is really all in the certificate handling. there's no
RW> certainty when a message arrives that is signed and/or encrypted that you
RW> have the certificates in hand needed to deal with it. if you have them,
RW> then the rest of the work isn't so bad.

for me, this is the primary reason why s/mime is evil.  As soon as you
introduce certificates, this means you have to have a certificate
authority.  Self signed certificates don't even have a web of trust to
fall back on.

RW> good UI decisions here will determine whether or not S/MIME support would
RW> be useful or ignored.

I wholeheartedly agree.  And I'll also add that the best UI for
encryption is no UI.  Encryption and decryption should happen
automatically, invisibly, without any user intervention necessary.

The reason I say this is from watching companies desperately try to
institute good password practices only to create a run on post-it pads
every time password changes were mandated.  You could almost hear the
keyboard flipping over in unison as people "hid" their new password.

you are lucky to get people to type their password once on login and
if you make them logout every day, they get really annoyed.

Therefore, if you're going to try and get users to type a pass phrase
when they e-mail messages you are going to encounter the same UI
failure has existed with every encryption program since PGP was born.

And antispam system I'm working on (www.camram.org) will add some form
of 0 UI spam protection and encryption.  In a strict form camram
system, a mail message must have 1) a proof of work postage stamp, or
2) a signature from a key you have accepted.  in order to make this
work, public keys have no pass phrase, keys are passed automatically
in mail messages as part of the header and therefore must be
relatively small.  Keys are also considered disposable and can change
at anytime without revocation or any niceties.

before you start sputtering about the weaknesses, they are recognized
as part of the trade-off for a 0 UI system.  There are also manual
processes one can go through to strengthen the system such as out of
band communication of key signatures etc.

in any case, I think we both need access to the e-mail stream at the
same points.  On inbound e-mail, we need early access to the message,
on outbound e-mail we need late access to the message (before
transmission).

---eric




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Mahogany-Users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mahogany-users

Reply via email to