There's no room at all to accommodate this idea.
1. It's too late to add new features.
2. There's nothing that assures the time stamps are accurate. Any attempt
to introduce time stamps into a security protocol, now or in the future,
will bring substantial debate about the accuracy and integrity of the time.
3. Any group of users who wishes to include time stamps can do so.
The only challenge before us is whether we can avoid getting bogged down in
a debate on this unnecessary topic.
Steve
Steve, I'm not sure how it can be "too late", when we are talking about
advancing to standards track a document which we haven't yet seen, i.e., the
rewrite of the PEM/MIME spec for the non-MIME related stuff. Likewise, I would
hope that discussing an idea doesn't bog us up, down, or sideways. Ned, Amanda,
Jim, and others don't have to put down their coding pencils -- if and when we
decide to include such a construct in the specification, I'm sure that it
wouldn't take them more than about a hour to include it in their code.
I was not suggesting that the time stamp affixed by the user would necessarily
be accurate. The primary reason for including it is to avoid a replay attack,
and for that a random IV would be sufficient. However, a time stamp on a
document, even if somewhat inaccurate, places a document in a sequence, and it
is unlikely that the time would be wrong by more than an hour -- almost
certainly not a day. But if I receive a document that is dated in the future,
or is significantly stale, as a recipient I would certainly have cause to
wonder, and perhaps check a little further.
In your previous message you asked about assurance. Time stamps, certificate
expiration dates, certificate revocation mechanisms, and the option of
high-quality identification policies and binding of the name/identity to a
public key are the primary differentiators between a system that was supposed
to provide the option of high assurance (PEM) and a system like PGP (which
already includes a time stamp, as does PKCS if my memory serves me correctly).
Certainly individual users could add a time stamp if they think to do so, but
then there would be no consistant place to check for it. If I want to sign a
free-standing document, I guess I would have to add another MIME body part just
to contain the time, and if I had multiple body parts and wanted to keep track
of when the individual parts were signed (e.g., in an approval process), I
would ahve to keep adding more and more parts. This doesn't constitute an
architecture, in my opinion -- it is just a hodge-podge of ad hoc solutions.
We can and should do better.
Bob
--------------------------------
Robert R. Jueneman
GTE Laboratories
40 Sylvan Road
Waltham, MA 02254
Internet: Jueneman(_at_)gte(_dot_)com
FAX: 1-617-466-2603
Voice: 1-617-466-2820