On Sun, 12 Sep 2004, R. A. Hettinga wrote:

From: Adam Back <adam@cypherspace.org> Subject: Re: anonymous IP terminology (Re: [anonsec] Re: potential new IETF

At ZKS we had software to remail MIME mail to provide a pseudonymous email. But one gotcha is that mail clients include MIME boundary lines which are pseudo-random (purely to avoid string collision). If these random lines are generated with a non-cryptographic RNG it is quite likely that so called unlinkable mail would in fact be linkable because of this higher level protocol.

Wouldn't it be relatively easy to regenerate the MIME boundary strings on the level of the remailer, and filter the content of the headers? Various mail clients have various peculiarities, "fingerprints". Shouldn't the remailer be able to break the message down to individual data objects (subject, message text, attachments...) and then reassemble them back, in a sanitized way?