Are you looking for stuff about RL cryptography - or more crypto puzzles for people to solve?

Currently the common crypto algorithms run based on how extremely difficult it is to take a remainder from division and figure out the numbers used to derive it. I suppose if you don't know ANYTHING about the message prior (including a document signed by the sender's private key), you could be generating a list of random original messages as guesses. However, if the message is of sufficient length, it's not hard to eliminate 99% of them on the grounds that it simply isn't language. This isn't a situation where any original message is possible while trying to reverse/guess the encryption, so it's not a huge issue. The way cryptography works right now is, basically, we have a set of problems called NP, which means Not P, with P meaning, more or less, 'easy'. These problems are tremendously difficult and time-consuming to calculate. But doing that same problem the other way (i.e., work from the answer to find the question), it's P - very easy. (An example is above - X divided by Y has a remainder of 13. Solve for X and Y. Compare to 283 divided by 45 has a remainder of Z.) The problem EP brings up is that Quantum Computers (and, although it's not mentioned, biological computers) solve problems in a fundamentally different way. Many problems which are NP for a computer may be P for a biological computer (and, funny enough, vice versa). So once you have a quantum computer, the classic encryption methodology has to move on. You are ultimately going to have a few types of encryption in EP: P for everyone. Examples are Caesar Ciphers, letter substitions, foreign languages, etc. These can be cracked in trivial time. While currently the mystery can be one of 'guess the method' (is this message a letter substitution? Is it written backwards? Is it Cherokee?) in EP you're going to find that extremely ineffective unless the sender is so alien in mindset that nothing in our entire racial history would give us a clue as to what it means (for example, a message written in the Factor language is effectively encrypted and may be impossible to crack without further information). NP for computers, P for quantum computers. These are some sort of problem which, as I said above, takes a P amount of time to solve one way, but an NP amount of time to solve the other. Examples include the modulus (remainder) problems, pathing problems (like the travelling salesman problem), etc. If you read the wikipedia article on NP, this should give you some background, but you don't need to dive too deep. Likely, the majority of these can be cracked by quantum computers in P time. P for quantum computers, NP for normal computers. This encryption is just for a laugh. No one would really use it. NP for everyone. Because I don't understand quantum computers, I don't know what an NP quantum computer problem is. Apparently the authors of EP don't either. You could be the first! Go for it. An important note on the 'P' section above. Note that I said 'anything which cannot be predicted given the whole of our racial history'. Anything which you can create that does not tie into our racial history can create an effective algorithm in this category which will be effective. An example of this is a one-time pad. A one-time pad is when you have a super-long code. When you send me a message, you start it by saying "start with character 4,528 (or whatever) of our shared code'. You then run each number in the message through an XOR function with each number in the code. The result is effectively unbreakable - as long as the code is unpredictable. If our code is a book, or if it's intercepted in transit or so on, the bad guy spy can likely crack it. An example source for a code is recording the audio feedback on a microphone, or certain functions regarding heat, or the decay of radioactive materials, which are effectively unpredictable. The code then needs to be sent securely, likely by courier, to everyone you want to have access to it. It needs to be kept secure, and you need to make sure you never use the same part of the code twice. Quantum encryption in EP apparently may function very similarly to this. Again, it's a P problem. But unlike a one-time pad, the key can't be copied or determined from data analysis. Quantum encryption may also be an anti-tampering mechanism - if the message is opened, the quantum bits no longer match, and both parties know a third party has read it. Not the same as encryption, but still a useful tool. Hope that helps.

I guess another thing to consider is the life time of the data being protected. While a 128-bit AES can be broken in 24 hours, if the protected data only have value for a manner of minutes its good enough.

Quote:

Do you want real life existing crypto algorithms (like the algorithm to AES) or are you looking more for theoretical crypto (like quantum crypto)? Note that current crypto (no matter how good it is) has a shelf life of about 10 years max. After that it is effectively useless (no one relies on DES any more as an example). Since EP is at the very least 70 years beyond today, our current crypto would be as effective as Ceasar Crypto is today.

As a noted exception, there are one-time pads which are guaranteed unbreakable as long as the pad isn't reused. However, there are practical limitations in safely transporting the pad to the entity with whom you'd like to securely communicate. Escorting a shipment of one-time pads might even make for a decent EP mission.

nezumi.hebereke wrote:

In fact, the current quantum computers take quite a good deal of time just to do basic functions like arithmetic. However, yes, it would solve in P time rather than NP, which is an order of magnitude of difference.

No, NP is *WAY* more than an order of magnitude (compare the time it takes to solve the travelling salesman problem (N!) and bubble-sort a list (N^2) - you can get as many orders of magnitude of difference as you want by using a large N). However, quantum computers just give you an exponential speedup (on some problems). They do not turn NP into P strictly speaking (see the writings of Scott Aaronson, "Great ideas in computer science", a bit down in the right column of http://www.scottaaronson.com/blog/ ).

nezumi.hebereke wrote:

Currently the common crypto algorithms run based on how extremely difficult it is to take a remainder from division and figure out the numbers used to derive it.

Almost but not quite. Factoring a composite value formed from the multiplication of very large relatively prime numbers is more accurate.

nezumi.hebereke wrote:

Quantum encryption in EP apparently may function very similarly to this. Again, it's a P problem. But unlike a one-time pad, the key can't be copied or determined from data analysis. Quantum encryption may also be an anti-tampering mechanism - if the message is opened, the quantum bits no longer match, and both parties know a third party has read it. Not the same as encryption, but still a useful tool.

As a bonus, it would also likely corrupt the cyphertext,rendering it useless to anyone.

For foundational stuff, I'd go with Bennett and Brassard's 1984 paper on the subject, or Brassard, Crepeau, Jozsa, and Langois' 1993 paper "A Quantum Bit Commitment Scheme Provably Unbreakable by both Parties". Both should be available through your local public or university library.