> The archives (when ready) will be stored on a machine for which
> accesses are not logged. It may also allow SSL access (with strong
> encryption, obviously), which will include making available dummy files
> of various sizes so that it's not possible to prove (based on traffic
> analysis) exactly what was downloaded ("Crypto? Certainly not, I was
> downloading this paper on the history of Ethiopian pottery in
> 4000BC").
Just don't forget to make both the filenames and the file sizes
identical in length.
I've always wanted to set up some secret-sharing filesystem where
you have to download multiple "shares" to reconstruct the data.
But other combinations of those exact same shares give other data.
E.g. file1 XOR file2 is smut1.jpg
file3 XOR file4 is smut2.jpg
... etc.
However, file(2*n) XOR file(2*n+1) is... interesting.
If you break files up into lots of little blocks and do this with the
blocks and then put the blocks on different servers, it gets awfully
hard to say who's breaking which laws about what kind of forbidden
data.
It's kind of an idle thought, but amusing occasionally.
--
-Colin
