RE: Context sensitive encryption
Michael Elkins (elkins@aero.org)
Fri, 16 Feb 1996 12:57:25 -0800
On Feb 16, "Edward A. Russell" <erussell@ftp.com> wrote:
> That makes my life absolutely miserable.
>
> I am operating in a Win95 MAPI environment where I have no access to
> the transports (I have no idea what transports I might be running over).
> Any mime conversion is done by the transport. But I want to handle PGP
> up in my application so the messages STAY encrypted in the message
> store until the user deletes them.
>
> If I decrypt a message and I now have a mime encoded object I won't
> know what to do with it (that is, again, all the mime stuff was already
> down in the transport).
This is a problem for clients that support IMAP as well. I found this out
after I implemented a Unix IMAP client and I tried to implement the PGP/MIME
draft and couldn't do it, for two reasons: (1) I could not verify a
signature because I did not have access to the "content" headers, and (2)
my client did no MIME parsing.
> I wish PGP "objects" were simply wrapped or whatever as MIME objects.
> That way, the transport delivers to me PGP objects (as attachments or
> message body) and I can deal with it under user control.
>
> Someone said last week (and I may have gotten it wrong) that part of
> the reason for your scheme is that you cannot encrypt or sign things
> without converting it to 7-bit ascii, or 64 bit radix or whatever. I still don't
> understand that.
Just a minor clarification: it is _OK_ to encrypt 8-bit (or binary) data
without conversion to a 7-bit format (Q-P or B64) so long as you are not
signing it. The reason for conversion to 7-bit in the case of signatures
is so that you can still transmit data across a 7-bit SMTP infrastructure if
you decide to strip off the encryption "layer" and still retain a valid
signature.
> I am taking binary files, sign/encrypting them or signing
> them using a seperate signature file, then MIME encoding them and
> sending them off. When I receive them, they are de-MIMED and I have
> either the encrypted binary which decrypts fine, or the seperate
> signature which validates the binary file just fine. I can PGP binary
> objects safely and I can MIME binary objects safely afterwards. Where's
> the problem. And if there is none, why do I have to have MIME objects
> inside of PGP objects?
The _big_ problem with doing it that way is that any observer knows what
type of data is inside the encrypted block. It is absolutely essential
that the content-type be hidden except to the recipient of the encrypted
data. The issue is not that PGP can't do it, but that it's a security
hole to do it that way.
me
--
Michael Elkins <elkins@aero.org> http://www.cs.hmc.edu/~me/index.html
PGP mail preferred. Key availible via web or 'finger -l me@cs.hmc.edu'
Key fingerprint = EB B1 68 32 3F B5 54 F9 6C AF 4E 94 5A EB 90 EC
"I could be wasting my time more productively than this." --me