unforum

[wintermute_au]

ok so I've gone through the string from BrianEnigma's file, replacing various numbers -- I came up with 26 different double-digit characters, plus 5 single digit (1/3/5/7/9) -- the attached text file is the result of that -- note that the choice of letters means nothing, but someone might be able to do some substitution work with it...

I also converted * characters to spaces and stripped out the +s....

EDIT: I recounted and I'm actually getting 8 single-digit numbers (odds plus 2/4/ (bwahaha 2^1/2/3)

[Leeravitz]

Just to get the basic feeling on this: do we figure it'll be ultimately possible to crack this through brute force??

I don't how literally we need to take the notion that it's been encoded using earth compatible tech, in which case perhaps only applying the right *technical* decryption to it will yield up anything.

Clearly, getting a transcript of the complete string might be an initially useful (if tedious) job.

Then, I guess we can basically say, even at this stage, that the character set in use is highly limited - numbers 1 to 9 in various combinations, and the symbols *, +, ^, <, >, @ etc. and the prominence of a single letter (if, indeed, it is a letter symbol, rather than a pure geometric shape), which is rendered 'v'.

The logical consensus (so far) would seem to be that: the numbers are standing for the encoded message, and the symbols represent variously:

<> direction in which the string should be read
* + Potential delimiters of some sort - indicating end of word?/sentence string - although peculiarly we sometimes get more than one such symbol in conjunction
@ Probably the 'full stop' at the very end of the message?

Given that the numeric strings are rarely very many characters long (that is, if we accept that, say, a * or + marks where one numeric string ends and another begins), then I guess the assumption has to be that the encoding has been broken up in a way which does not correspond to the equivalent English language spacing pattern for the message - that might render a frequency analysis, as such, fairly pointless - although it is clear, even with cursory examination, that 9, for instance, has an incredibly high frequency of usage throughout the message.

I am not certain whether it might not prove useful to separate out the lower portion of the message (that which does not take on the 'concrete' form - I believe that's the technical term for using the symbols to make a pattern) from the portion shaped like a Cube. Simply because this looks more interesting we're focussing effort on it, but frankly, the strings in the lower section are much easier to transcribe, and if they can be cracked, then the Cube will yield up its secrets...

Of course, that does assume that the 'break' between Cube and 'underwriting' is purely an aesthetic inclusion to make the Cube shape stand out, and further break up the overall message (who knows? It could be an image like an 'I Ching' hexagram with meaning for the Recons - a complete picture of the Cube resting on the earth that it had come to rest upon, surrounded by the thin lines of space from which it descended??). On the other hand, if the 'break' indicates a switch in key usage, then solving one bit of the puzzle won't yield up the other half of the answer...

But it seems to me that the big question at the moment is what sort of basic encrypting pattern are we likely to be looking at here, without worrying about what it actually encodes. Given that we have only (primarily) the numerals in combination to stand for whatever is encoded, what do we think we're looking at here? Have many times do we think the thing has been encrypted? If, for instance, the numbers do turn out to stand for letters, then will those letters themselves need further decryption in accordance with shifts etc? Given that we clearly fall short of any simple substitution alphabet, are we to assume that numbers 1 - 9 have been used to stand for multiple letters??

Those bloody Recons - cleverer than we thought...

[And what sort of tech. expert is Kurt exactly, if he has to hand this over to us?].

I don't know - first you get a puzzle like the keycode which someone solves pretty quickly, but still makes minimal sense even when it's been done - now this, which might take a bleeding long time to break...

Oh, I get it. The intention is: we'll be trying to work this one out for the next two and a half years while they get the cards ready for distribution...

Where's an Engima rotor when you need one??

[Zedix]

Well, it seems the + could delimit words. Single digits could either be vowels, or more likely, punctuation. * seems to delimit figures. In which case you'd have 26 double digit figures seperated with *, words seperated with +, and punctuation, such as period, comma, exclamation mark, and question mark, then either quotes, colon, or semi-colon (5 of them), but those would be fairly easy to see at the end which is which.

I haven't actually counted the double digit figures, but if there really is 26, then it can't be a coincidence and we really can substitute them for letters (although it's definitely not ASCII).

[MasterCheese]

[Leeravitz]

Good call, Zedix. If the pattern really is anything like that, then it might be a damn sight easier to get something out of it then we'd been assuming...

[wintermute_au]

ok this thing is scaring me more and more every time i look at it and i need some sleep

but i will point out a couple things... I originally was assuming that the '*' character delimited 'letters' and the + delimited words...

however, I'm noticing a few other patterns where the *+ come in to play.... some examples:
"*1+" and "*5+" and "*39*91+91+*01+"

so i'm still not sure what to do with our string -- there are definitely 26 double-digit characters, but the 8 leftover singles and the *+ still don't have a defined role.... so keep brainstorming people

[Zedix]

I've also been looking at the ones who are isolated. If you use + as word delimiters, there's only so many letters which can be single right? I think if you take 'a' and 'i' for example, they are the only letters who can be both in a word, and alone. Numbers who are found alone and *not* in a word could be punctuation.

[QBKooky]

Here's the beginnings of my speculation on http://forums.unfiction.com/forums/files/recontext5_196.txt (the attached file from the top of this page).

J = period.
J is the last character of the file, and also only appears at the end of words.

Also, the letter C. C only appears at the ends of words, too, or on it's own, in often-repeated phrases like "C 1", "C 3", C 5", etc.

Other common "words": "OC" (53 times), "EC" (14 times), "AB1" (19 times), "N9Q1" (23 times), "K9K1" (24 times).

[Leeravitz]

Two new things:

1. If we've got the 26 double - digit characters, is it worth doing a quick frequency analysis on those alone, and seeing what sort of results we come up with?

2. I don't know how much we need to read into this, but...assuming that this isn't ASCII (and it wouldn't seem to be), then we're moving further away from Violet's statement that the encoding was done using some specifically Earth - based technique. What puzzles me is that, surely, the Perplexians have plenty of native cryptographic technique to draw upon, so the fact that Violet implies that the Recons. may have been using some kind of algorithim that would be familiar to Earth people (but, by implication, less well known to Perplexians, is odd). I think my immediate assumption on reading this was that the encoding would somehow bear the mark of some technique or other that was familiar historically in this world. If it is simply a code produced sui generis, then this makes Violet's/Kurt's statement strange, because then the Recons. could just as easily have been drawing upon encoding techniques that they were simply making up. It's as if Violet has handed the puzzle to us because she imagines that Earth people will be able to see more familiar usages within it. But, who knows? Maybe I read the (brief) entry slightly wrong - and, well, Violet's wrong - footed us a number of times before. I'll keep on thinking...

[Leeravitz]

Actually, I'll just revisit some of my previous comments

Ignoring entirely for the moment the actual decoding of the message, there are two further interesting points that I think emerge from what we've been given:

1. The fact that this type of material can be found encoded on a key (and, let us recall, this is an antique key - over 40 years old) may tell us something more about what keys are, and how they operate. That may not be a problem we wish to deal with in this forum, but it seems interesting...especially if the coding is reminiscent of, say, how we on Earth would have coded in the early 80's etc.

2. Looking carefully at what Violet says, she states that Kurt has told her that he thinks it looks like the thing has been encoded using Earth technology. I don't actually know what that *means*, when I think about it. The encoding has been done on a piece of Perplexian tech, I assume, and yet, Kurt thinks Earth technology has been used in the process. Does he mean, as I hypothesised above, that he suspects the algorithim used is Earth - specific (if that's possible?)? If not, then what is he suggesting? It could be that Kurt's wrong, it could be that Violet's phrasing things badly again (and she meant to say something like 'Kurt thinks the code usage is one that's familiar on Earth'?).

It also makes you wonder about: a) why the Recons would encode using Earth technology [unless it's simply assumed to be more secure, because harder for Perplexians to crack] b) whether this gives the lie to the Recons removing themselves from all contact with the outside world (how else do they even know what Earth technology is like?) etc.

Equally, given that the key concerned has existed for 40 years or thereabouts, it could be the case that the encoded entry isn't even something recent...

I realise little of that helps us further with the actual decryption, but it all looks pretty peculiar from where I'm sitting (still, I never thought the Recons would rig a door password in reference to astrological formulations, so what do I know?)

[Axys Denyed]

Just because he hasn't posted it himself yet, BriEnigma worked up a little number frequence chart. Here is the link

[Wainamoinen]

*Edited because it was a stupid idea*

[leakingpen]

no... they wouldnt do that...

do you think the * and + are dead on? it could be godelized....
btw, theres grouping repetition. for example, near the top? you have 99*3*85+* seperated by directionals. and it repeats a few times. btw, 99*3*85 equals 25245, which is one of the "perfect" numbers, in which the sum of all its factors equals it. i checked other groupings, they repeat as well, but none of the others are perfects or primes that ive seen.

[firefox]

so someone in chat (you know who you are you special person you) suggested thats its really: Befunge
that should set the trail ablazing.. mind you chat is ahead by 30mins now, and Brien and jess should post what theyve found soonishly.

converting http://www.quirkyacuity.com/reconfile.txt to reconfile.bf (just changing extension)

and using: http://quadium.net/funge/tbc/ to convert the code to c

i get the output of: useless, good job on the solve hawk.


dont have time to do much more on this, examing in a few hours. good puzzle tho

[BrianEnigma]

Someone in IRC (sorry, I am not sure who to credit this to) pointed us to a thing called Befunge: http://en.wikipedia.org/wiki/Befunge

This looks like it takes ASCII, like our cube artwork, as input, then runs it as a program. Basically, the arrows define a path (although there are other symbols, not present in our cube art, that define other ways of moving around), the numbers, +, and * define RPN functionality on single-digit numbers. Nobody could get the Befunge example to work, as the input was too large, but taking the data we got from walking through the path into a simple RPN program (attached for those that care), I am ending up with a stack of numbers on the output. Mind you, I am not 100% sure my RPN program is functioning bug-free, as I have not had the time to manually spot-check values, but my results can be found here. The first number is the deepest one on the stack, the last one is the topmost value on the stack.

[b]EDITED[b] to add: Number frequencies are here. Again, since I have not had time to double-check this RPN work, please take it with a grain of salt.