[Remops] Why is 4096 bit Mixmaster consider 'experimental'?

[Tom Ritter]

That's the mixminion codebase, but not the sphinx packet format example.
 I'm not sure where it's canonical link is either, but I've uploaded my
copy to https://ritter.vg/misc/stuff/Sphinx-0.8.tar.gz I've tried to
understand it 2 or 3 times - still working on it.

Been watching this with a lot of interest.  I don't want to pile on more
work suggestions, but I did talk with people about Mixmaster and Mixminion
a few years, and I thought Mixminion was a more maintainable code base to
work on going forward. This was a todo list for Mixminion that came out of
it: https://github.com/nmathewson/mixminion/blob/master/TODO#L1

I think one of the biggest wins for the lowest time investment that could
be done with Mixmaster would be requiring StartTLS with DHE ciphersuites
between nodes, logging certs seen/changes (to detect MitM), and ideally
doing key pinning (to prevent MitM).  weasel has a lot of mixmaster scripts
that do interesting polling that haven't run in about a year and a half,
but back then, this was StartTLS support: http://www.noreply.org/tls/

-tom


On 21 March 2014 18:02, <lists at notatla.org.uk> wrote:

> > I think starting from scratch with something based on Sphinx[1] would be
> > a good option.  There's a reference implementation written in Python
> > but, despite significant Googling, I can't find it today.
>
> mixminion-dev at seul.org
> http://mixminion.net/
> https://github.com/nmathewson/mixminion
> _______________________________________________
> Remops mailing list
> Remops at lists.mixmin.net
> http://lists.mixmin.net/mailman/listinfo/remops
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mixmin.net/pipermail/remops/attachments/20140322/575d01cd/attachment.html>