# Using IF to communicate symmetric keys

Imagine you have to email a encrypted super-secret message. You don’t trust RSA anymore because maybe some quantum computer in NSA’s compounds in breaking keys faster that Emily Short can write new IFs (pun intended). You decide to stick with basic symmetric algorithms, maybe one of the offspring of DES. But how do you communicate keys?
Easy. You use IF.
You employ a IF author to create a large game with numerous room. In one of the room, you introduce a new verb that is actually a code word. This codeword is generated using a complex mathematical function and you can easily communicate it using some pdf on mathematical treatise, uploaded on scribd. Then you upload the IF on ifarchives. You send a email to your correspondent- something like “The Red Room of Horror house looks good on next friday”. This message, most probably, raises no Bayesian alarms. Your correspondent downloads the game “Horror House”, goes in the “Red room” and put the codeword (which he generated by putting friday’s date in the function) and is rewarded by (in a textdump) a perfect pseudo-random key to decrypt the message.
And while the spooks were going after every e-mail, software, gif, txts with ending whitespaces (aka SNOW); you used something they maybe don’t even know exists- IF. Ofcourse, the idea can be polished but it’s interesting nevertheless, right?
PS. And maybe a new concept for a new IF! Hey zarf, waddya say?

I say that the NSA can snoop on this forum a lot more easily than it can snoop on RSA-encoded messages.

And then what? They reverse-engineer all the IFs ever made? I know they are retards but not so much. And anyway, it was just an idea (most probably, it has been implemented already)

Using IF as a hackery method to get back at The Man is an idea at least thirteen years old. See Breaking the Code, an entry into IF Comp 2000.

My god, Paul Panks was a super spy.

Anyone sufficiently creative can come up with good espionage solutions. I am now tempted to write an intfiction that, when you solve the puzzle, leads you to real-world co-ordinates where treasure has been buried.

There have been people on here before talking about using IF as an element of geocaching challenges, too.

This is all coming around to Domark’s Eureka! and the Masquerade challenge that spawned it…

A large community-aided project, on the lines of Halo’s I Love Bees, can be a good idea too. Or maybe, instead of leading to coordinates, it may lead to URLs.

Guys, snooping isn’t tricked by formats. Among the first thing you learn as a computer user, is that if you save a txt as a jpg, the computer will think it’s a jpg while the contents is the same. Snooping looks at streams of data, and I bet that encrypted textdumps won’t be any more secret and encrypted, just because they’re listed somewhere in an IF file. If you have some kind of correlation maps in an IF, then you could possibly convey some abstract ideas where the meaning is already clear to both parties, but that’s basically it. You’re better off using simple codewords to avoid detection.

Sending pseudo-random keys in compiled z-code and sending encrypted stream in IF are TWO DIFFERENT things. The problem with symmetrical encryption systems is secure communication of keys, not sending encrypted message. A large textdump, which actually is a passphrase, can be used to communicate keys easily; that’s all I was saying.

I think he was saying that the information would be hidden inside the game, printed when certain conditions were met. And one interesting “feature” of the Z-machine is that the text isn’t encoded in normal ASCII. Try opening an Infocom game with a hex editor and you’ll see. I have no idea why it was designed that way, but in this case it would actually be an advantage.

EDIT: Cross-post!

Well, if you’re not a person of interest, a lot of things can work out. It’s a fleeting scale between getting away with shouting things in the street, to having your computer keylogged.
You are basically talking about hiding a message in an “innocent” file, like Al Qaeda did with picture watermarks. (Since the communication got known, it was easy to decipher the watermarks.) I doubt it matters if Z-machine code doesn’t use ASCII, because it’s most likely a basic substitution cipher, and if you don’t trust RSA encryption because of huge machines trying to get at your message, then you might not trust a cipher that it just takes an average person with a person above 100 IQ to crack.
…but if you’re not paranoid, it’s clever.
However, if you’re paranoid, you might want to try something called “elliptic curve crytography”, which makes RSA look plain stupid by comparison. Unless you believe the rumours of NSA possessing backdoors to those algorithms, there’s no computer that will ever be able to crack them.

It’s called Compiling. And that’s why I proposed the scheme in the first place.

Curve Based Encrypytion is as simple to break as RSA using a Quantum Computer (which, as the rumors state, NSA is trying to make).

Andreas: That’s very true, it is effectively a simple substitution cipher (a different character encoding called ZSCII). It would be easy to find if you were looking for it–it’s not very good encryption. But it could be very good steganography: rather than making the message hard to decode, it’s not obvious that there’s a message there to decode at all (like with the watermarks in the pictures).

Naman: True, but except for a one-time pad I don’t know of any form of classical encryption that can stand up to a quantum computer.

Goppa polynomial based encryption, lattice based encryption, etc.