On the Run Options

[Ophis]

Because, Techno's are an urban myth...

a fact missing from the book I know, but hey, someone from the freelancers mentioned this a while back. Synner I think.

[Dashifen]

And, if I remember correctly, the encryption lasts only as long as the sprite, so if it's not registered, the the encryption is temporary. Plus, if the sprite is derezzed in cybercombat, then the file is irreparably destroyed. One problem with data is that if it's important enough to encrypt, it's also usually necessary to be able to decrypt it someday and use that information.

[Cain]

Oh, come on. Ultraviolet nodes are a myth as well, and plenty of corps invest in those. So are Deltaware clinics. And AI's. The fact that immortal elves are myths didn't stop Renraku from working with one. Corps will follow any myth, if it stands a chance of paying off.

The new otaku are gone over in detail in System Failure, and the corps have *got* to have eyes on the data havens. There's absolutely no reason why the megas wouldn't be intensely researching otaku abilities.

[Butterblume]

[Ophis]

I think someone already had that idea.

The first big plot book is called Emergance and will apparently have perception altering ramifications for the matrix...

I just wonder what will emerge.

[neko128]

[Cain]

[James McMurray]

Not if you use the aforementioned cosmic radiation method. The key doesn't have to be someting rememberable, or even intelligible.

[Cain]

As Butterblume also pointed out, it's impossible to prove that anything's perfectly random. Even that method has flaws, and can sometimes produce patterns that can be cracked. It's a lot more secure than most, I'll grant, but you're still running a greater risk with a longer key.

[Azathfeld]

[Ranneko]

Computers are able to produce psuedo-random numbers that are random enough today.

The tricky thing is of course while it is impossible to break a message if you use a one-time pad once, if you make the mistake of reusing a one-time pad it becomes trivial to break.

[hyzmarca]

With enough computing power anything can be decrypted by brute force, even a one time pad. One just has to keep trying keys untill you get a message that makes sense.

The difference is that the one time pad requres significantly more computing power because the decrypting program must be able to determine if a result makes sense or not.

Presuming that the message itself is writen in a known language and follows standard syntax rules an AI could brute force it. It isn't foolproof. With countless trillions of possibile keys it is possible for an incorrect key to produce a grammerically and contextually correct sentence.

Pseudo-random numbers produced by computers aren't random at all, however. The algorythims ar eentirely deterministic and thus it is possible to know exactly which number will be chosen in what order with enough information.

[James McMurray]

No, the AI could brute force out a message of the right length that makes sense. There's absolutely no way to ensure that the message you banged on keys to get is the actual message.

XX YY ZZZ2 means "Go Up Yes." Your code breaker is just as likely to stop at "Is it bad?" because it makes sense, and will probably bypass go up yes altogether.

That brings up another method of encryption: vague words and slang. A message encrypted to mean "chill it out." will mean totally difference things to a Springer guest and a chemist.

[Shrike30]

Or you could do what a bunch of the orignial 3133+15+5 (eleetists) were actually trying to accomplish when l33t got invented, and use a total hash of characters that actually means something to the person on the recieving end. Or do what people did in CP2020, and "invent" a language (and pass out chips for it to the people you want to be able to speak it) that consists of real words assigned cosmic-radiation-randomized meanings ("Fish table or yuck!" translates to "I don't want that!"). Or you could nest cyphers within cyphers... it's not like we don't have enough computing power on hand to run a simple message through 20-30 cyphers instead of just one. Or you could send a digital waveform of someone speaking the message, rather than just the characters... we've got the storage space and the processor power to make one-time pads large enough to allow this.

The point here is not to make an unbreakable code. The point is to make one that's such a pain in the ass to decode (even for a dedicated SR4 AI with way too much time on it's hands), and takes so damn LONG to decode that nobody tries, or at least they don't succeed until years after it's too late.

Or you could just run Decrypt on it :P

[Azathfeld]

[James McMurray]

We know 2070 isn't that time, because apparently On The Run has an unbreakable code in it.

[Cain]

[b1ffov3rfl0w]

[Dranem]

[fistandantilus4....]

I'm wondering how they're supposed to be making 'new'technomancers, since the existing ones were all people that were in the system during the crash. A bit different from Otaku. I also don't see how the otaku are supposed to automatically become technomancers when they 'grow up'. Hopefully this will all be explained in Unwired. Unless I simply missed something in SR4 or SF that covers that.

[Crusher Bob]

Once you have your random numbers (for your OTP generated) you run statistical tools on them and look for patterns (in case somethnig went wrong with your random number generator). If you find patterns in the OTP, you throw it away and use the next one (or get a new entropy source). However, random numbers based on radioactive deacy do appear to be completely random, you just have to make sure that your data dosen't randomly have patterns in it.

Of course, the failures of encryption are almost never due to actual breaks in the cypher text, but instead some failure of the human elements involved. In general, much 'cheaper' encryption is also esentally unbreakable, so OTP are hardly used for anything.


Of course in SR, all non-OTP encryption seems to be trivially breakable, so your only option is to go with the expensive and tedious OTP.

In general there are three 'strengths' of encryption that anyone will want to use:

Tactical/short term:
decryption time: hours/days where the decryption of the information is only useful for a very shotr amount of time. For example, if the team's tactical radios can becrypted only after 4 hours of work, then the fact that 4 hours ago, you said 'shoot the guy on the left' may not be that helpful. Notice that the team must practice comm discipline, and not discuss anything of 'in the open' or of significance:

Example of good como use:
Infil team to fire support team: "Egressing via route bravo, we will rejoin you at rally point Silver"

So, notice that things that might have some long term significance, such as the route the team will be using, and where the rally point is are 'encrypted' with 'idiot code', and thus are unavailable to someone breaking the comms encryption.

Long Term:
This is the stuff that you deal with on a day to day basis, it might be harmful if revealed, even after everyone involved is dead. You'll want the decryption times to be in the hundreds of years (remember to future proof your encrpytion schemes).

For all time:
This is the stuff that can never get out. It's what you use OTP for.

[mdynna]

Getting back to On the Run... here's a thought I had. Why can't the PC's totally decrypt the message? What would be so terrible about it? They'd try and sell it? The adventure already makes provisions for what would happen to the PC's should they attempt that. Otherwise, its just for their "private listening". I'm thinking this to avoid the whole mess of the "unbreakable code".

[neko128]

[neko128]

[neko128]