future proof file encryption

[Robert J. Hansen]

Sven Radde wrote:
> Imagine the session-key part of an OpenPGP message be destroyed.
> Commonly, this will be far less than 1% of the actual data, but even
> with 99% intact, you won't have a chance of recovering *anything* from it.

Err.  What?

With a 256-bit cipher, if you're missing 3 bits, there are only eight
possible keys.  This is not an obstacle.

> With a single images.zip.gpg file, a bit flip may mean that the whole
> archive is unreadable (which is the worst case... no idea what an
> average case might look like).

The moral of the story is not to avoid encrypting your backups, but to
keep multiple copies of your backed-up data.