5 ms·

Clearly, just use a standard solution (PGP). Don't do this with a weak passphrase, and use key-strengthening (bcrypt etc.) just in case. Assange is pretty good

Clearly, just use a standard solution (PGP). Don't do this with a weak passphrase, and use key-strengthening (bcrypt etc.) just in case. Assange is pretty good at crypto, so I'd expect Wikileaks to broadly get this right. Also, they already got in trouble for reusing a password (http://boingboing.net/2011/08/31/wikileaks-guardian-journalist-negligently-published-password-to-unredacted-cables.html http://boingboing.net/2011/08/31/wikileaks-guardian-journali...), so I'm pretty sure they won't do that again...

raverbashing 13y ago
No, the key strengthening is probably built-in. Also, most likely it doesn't use a passphrase, but a big generated gpg key (with a passphrase protecting it, of course)
- JoachimSchipper 13y ago
  At least for symmetric encryption, gpg defaults (defaulted?) to SHA-1'ing the passphrase 2^16 times, which is not a very large number - http://passwords12.at.ifi.uio.no/Jeremi_Gosney_Password_Cracking_HPC_Passwords12.pdf http://passwords12.at.ifi.uio.no/Jeremi_Gosney_Password_Crac... [pdf] talks of ~2^36 operations per second. This isn't the worst thing ever, but using GPG for passphrase-protected encryption is not as strong as you might expect - just use a slightly longer/better one. (Obviously, don't switch to another tool if you can't evaluate it in depth - GPG isn't perfect, but picking a random other tool won't be an improvement.)
stefantalpalaru 13y ago
> Assange is pretty good at crypto Not really. He gave a moron the password to a publicly distributed insurance file instead of making a new encrypted file just for him.
- JoachimSchipper 13y ago
  Are you aware of http://en.wikipedia.org/wiki/Rubberhose_(file_system) http://en.wikipedia.org/wiki/Rubberhose_(file_system)?
  - stefantalpalaru 13y ago
    No, I wasn't aware of it. The concept looks similar to TrueCrypt's plausible deniability.
    - JoachimSchipper 13y ago
      Yes, basically; RubberhoseFS helped bring the idea into more-mainstream tools.
    - trevelyan 13y ago
      I think he is suggesting you read up on the authors.
      - stefantalpalaru 13y ago
        I've done that, but the project is long dead and it doesn't really showcase the technical chops of its authors. It's more about good intentions.

raverbashing 13y ago

No, the key strengthening is probably built-in. Also, most likely it doesn't use a passphrase, but a big generated gpg key (with a passphrase protecting it, of course)

JoachimSchipper 13y ago
At least for symmetric encryption, gpg defaults (defaulted?) to SHA-1'ing the passphrase 2^16 times, which is not a very large number - http://passwords12.at.ifi.uio.no/Jeremi_Gosney_Password_Cracking_HPC_Passwords12.pdf http://passwords12.at.ifi.uio.no/Jeremi_Gosney_Password_Crac... [pdf] talks of ~2^36 operations per second. This isn't the worst thing ever, but using GPG for passphrase-protected encryption is not as strong as you might expect - just use a slightly longer/better one. (Obviously, don't switch to another tool if you can't evaluate it in depth - GPG isn't perfect, but picking a random other tool won't be an improvement.)

stefantalpalaru 13y ago

> Assange is pretty good at crypto Not really. He gave a moron the password to a publicly distributed insurance file instead of making a new encrypted file just for him.

JoachimSchipper 13y ago
Are you aware of http://en.wikipedia.org/wiki/Rubberhose_(file_system) http://en.wikipedia.org/wiki/Rubberhose_(file_system)?
- stefantalpalaru 13y ago
  No, I wasn't aware of it. The concept looks similar to TrueCrypt's plausible deniability.
  - JoachimSchipper 13y ago
    Yes, basically; RubberhoseFS helped bring the idea into more-mainstream tools.
  - trevelyan 13y ago
    I think he is suggesting you read up on the authors.
    - stefantalpalaru 13y ago
      I've done that, but the project is long dead and it doesn't really showcase the technical chops of its authors. It's more about good intentions.