[liberationtech] OneTime 2.0 (beta): one-time pad system.

[Karl Fogel]

Look at this as a chance to test the hypothesis that one-time pad
systems aren't very useable in practice... :-)

OneTime 2.0-beta is ready for review and testing, as threatened [1].  See

  http://red-bean.com/onetime/

for download information, etc.

OneTime is meant to make one-time pad encryption useable -- or as
useable as it can be, anyway.  You have to supply the pads (make sure
you get truly random data), but OneTime tries to handle the bookkeeping
to keep track of which parts of which pads have been used up.

Version 1.x has been out for some years, and is packaged in Debian.
Naturally it's hard to know who's been using it, though I've heard of a
few instances anecdotally.  Although this release changes the encryption
output format [2], OneTime 2.x can still read 1.x files, and I tried to
make the upgrade path as automatic as possible.  Most users should be
able to just start using 2.x and have the only noticeable consequence be
that their interlocutors also need to upgrade.

I'm calling it "beta" until it's gotten some review, and won't update
the Debian package until there's been some time for review.  One-time
pads are so simple that implementing them in software doesn't count as
the sin of "writing your own crypto", IMHO -- though sharp-eyed
reviewers could certainly prove me wrong :-).  The point of OneTime is
just to provide a UI and a standard format.  But that doesn't mean it
can't have bugs, of course; please note that neither OpenITP nor any
other organization I'm affiliated with is responsible for such bugs.

I'll be offline for a few weeks on vacation starting around now, so
please just post comments as issues in GitHub (or Gitorious, if you
prefer; it's hosted in both places and I'll check both).  And if you're
a user, I'll be happy to exchange pads with you the next time we see
each other at a conference.

Best,
-Karl

[1] https://mailman.stanford.edu/pipermail/liberationtech/2013-July/009981.html

[2] It's a long story why the output format needed upgrading -- it has
    to do with making more efficient use of pad space; 2.x is noticeably
    better than 1.x in this regard, except for plaintexts that were
    already compressed.