[Remops] A new Mixmaster is in the works

[lists at notatla.org.uk]

> Today Lance Cottrell (the original author of Mixmaster) myself, (chief
> Mixmaster Alpha tester and contributor) and Anthony Papillion (Cypher)
> had a meeting on Jitsi to discuss the current status of Mixmaster and
> remailers in general.
> 
> We have decided that we are going to investigate updating to a new
> Mixmaster that will fill the weaknesses and issues that are present in
> the current version of Mixmaster (Mixmaster 3.0).
> 
> The first issue will address the 1024-bit key length. We will be
> looking at the development code currently available on the Internet.
> The first thing we will do is to implement a 4096-bit key structure.
> 
> More features are under discussion. We would like your input on what
> features you would like to see. Please keep in mind that we are
> looking at functionality first. Secondly we are looking at introducing
> a desktop GUI for major operating systems. Lastly we want to make
> using remailers and encryption as seamless as possible.
> 
> Your thoughts and suggestions are welcome but please remember that it
> is just the three of us so let's keep the "exotic" stuff to the end of
> the project.


I suggest looking at (and perhaps taking over) the neglected mixminion
project which includes such clever features as Single-Use-Reply-Blocks
and a "sphinx" packet format designed by well-known privacy-oriented
cryptographers.

I'd also try to keep it transport-agnostic. The dark mail alliance may
come up with something; or DH-connections between nodes may become more
appealing than traditional email (even STARTTLS email).

For a stop-gap we already have 4096-bit mixmaster.  Work on anything new
has to aim higher. Replacing a system with a fresh start is a relatively
rare chance to get something good and is not to be wasted.

I might possibly be able to contribute.