Talking in code...

This forum is for discussion related to the game.
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #21 (isolation #0) » Mon Jul 26, 2004 10:26 am

Post by mathcam »

Aside from the possible ethical questions, I don't see how people can claim that cryptrgaphy is usually bad. If it's used correctly (i.e. indecipherably), it can be an incredible boon to the town. As has been mentioned before, any time a player has information that is dangerous to reveal immediately but certainly needs to revealed
eventually
is a good candidate for encryption. Say I'm a role-name investigator and I figure myself to be a huge target if I come out, and I have some questionable (though not certainly evil) characters in the game. I can transmit this information to any player whose role name is definitely innocent by using the character's role name as the decryption key, by posting the encrypted message in the thread. Now that player has the information and use his/her discretion whether or not to reveal even after I die, and even better, it is unknown which player has that information. Keeping in mind that all my arguments are "in theory only", I should also point out that if this were to be a technique of yours, you should also occasionally do it when you have no information at all. This is so that the fact that you are encrypting something does not automatically give away that you have an information-gathering role.

There is absoultely no doubt in my mind that cryptography can be a tremendous boon to the town in many different scenarios.

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #24 (isolation #1) » Mon Jul 26, 2004 7:53 pm

Post by mathcam »

Only if the decryption is not trivial.
Agreed.
And you need some way to pass the key to only the correct person.
True, but often this is built in to the system. For example, in the role-investigator scenario, the investigator and the investigee already share a private key...the role name known to both players. There's no need in this scenario even for a key exchange. And like Norinel mentions, achieving a correct transfer is also quite easy (in theory).
And that person needs to be technically sopisticated enough to decipher the code.
Also true, but this basically reduces to being technically sophistacated enough to click on a hyperlink and paste in a large number.
There are still essentially unbreakable public key systems, right?
Yup, though the word "essentially" does require emphasis. First, any current cryptosystem (other than quantum cryptography) is susceptible to brute force given sufficient time and computing power. Second, no cryptosystem to date is proven to be unbreakable by some technique shorter than brute force. It's
possible
(though certainly unlikely) that RSA is easy to crack. But in short, "yes."

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #28 (isolation #2) » Tue Jul 27, 2004 10:00 am

Post by mathcam »

Malaprop wrote:Matcam -- one time pads, if not reused, have been proven to be unbreakable. Nature of the beast.
Okay, fair enough, but this isn't really a cryptographic protocol. It, in turn, has its own drawbacks, as Norinel points out.

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #32 (isolation #3) » Thu Jul 29, 2004 3:29 pm

Post by mathcam »

MMCL wrote:How about the box with two locks approach...

I have a message I want to send to mathcam (for example) so I use my procedure to encode it... I then display it in my post... mathcam then takes the encoded text and encodes it using his own procedure. He then displays the result. I then decode his result and post the message a 2nd time...
All mathcam needs to do now is decode the message one last time and he has the original text...
Depending on how you choose to "encode" in this example, this is called the Diffie-Hellman key exchange, used prominently both with RSA and with Elliptic Curve Cryptography. This is a decent protocol, but it hinges on the difficulty of cracking the encoding. If I can brute force the cracking of your encryption, then it doesn't matter how fancy of a protocol you use.

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #34 (isolation #4) » Mon Aug 09, 2004 4:42 am

Post by mathcam »

It does indeed necessitate commutativity, at least at first glance. There are non-commutative encryptions, but as you point out, these would need fancier protocols.

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #36 (isolation #5) » Tue Aug 10, 2004 10:53 am

Post by mathcam »

Hmm. I'm not sure if I'm being mocked or not.

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #40 (isolation #6) » Sun Aug 15, 2004 9:17 am

Post by mathcam »

Oh, now I get it. Some times...

Cam
User avatar
mathcam
mathcam
Captain Observant
User avatar
User avatar
mathcam
Captain Observant
Captain Observant
Posts: 6116
Joined: November 22, 2002

Post Post #46 (isolation #7) » Mon Jan 31, 2005 9:40 am

Post by mathcam »

It's fairly clearly (though admittedly indirectly) modquoting, which is prohibited (and, as DP points out, punishable by modkill) in almost all games on the site.

Cam

Return to “Mafia Discussion”