anonymous clearsigned messages (was 'possible new feature'/'security concerns)
Tue Nov 5 23:41:02 2002

Mindfuq wrote:
> I came accross the --throw-keyid switch today.  It's an interesting
> switch, in that it withholds the keyid of the receiver from the
> encrypted packets.  As a test, I tried clearsigning to see if it
> would withhold my keyid from the signature.
> Didn't happen.  But wouldn't that be an interesting feature?  It
> would protect a sender in the traffic analysis case described above.
> The receiver would either have to know who the message belongs to,
> or try all the public keys on their keyring.. but at least traffic
> analyzers would have one less peice of information.

you can do this now, if you really want to,
but it is a little *cumbersome*... ;)

[1]clearsign as you would normally
[2]use the throw-keyid switch to encrypt only the armored signature block
[3]replace the armored signature block with the encrypted signature block
[4]the intended receiver decrypts it, and pastes the decrypted signature block
back into the original signed message, in the correct place where the signature
block should be, and it verifies ok
{can send you some e-mail tests if you are interested}

but wouldn't it be simpler to just 'sign and encrypt' with throw-keyid switch,
as it is now?