Hmm…that does seem to imply the first few characters of the cipher (those following the paragraph marker) are perhaps a key to the rest of the message. I tried using the keywords I found at paragraph 56 in the book (the paragraph starting “Well, when Tom and me…”) as well as a few paragraphs before and after, in case I miscounted, but I can’t quite come up with anything. What’s confounding me is how those numbers are supposed to be used in the code.
It’s kind of NotPron-ish. Now THAT was an online puzzle. 
I don’t understand how the key word works. Could you explain it?
IF it’s a Vigenere, the normal way to do it is that you take the first letter of the cipher text and the first letter of the keyword and rotate the alphabet based on that. Then you take the second letter of the ciphertext and the second letter of the keyword, and rotate that much. Third letter, etc. With a Vigenere, the keyword repeats, so, if your keyword is 5 letters, when you get to the sixth letter, you go back to using the shift specified by the first letter. Another possibility is it’s not a Vigenere, but the entire length of the text starting at paragraph 56 is used to encode the message.
So, the typical way for a Vigenere would look like this:
HIMY NAME ISPU LYKA MELL - plain text
DOPE DOPE DOPE DOPE DOPE - keyword
------------------------
KWBC QOBI LGEY OMZE PSAP - cipher text
For the purposes of a standard Vigenere, an “A” in the keyword does not shift the text. A “B” shifts it 1 place. A “C” shifts it 2 places … A “Z” shifts it 25 places. So, following my text, “H+D” means shift “H” 3 letters forward. It becomes a K. “I+O” means shift 14 places, or “W.” “M+P” means shift 15 places. Once you get to “Z”, you just wrap around to the beginning of the alphabet, so “M+P” becomes “B.” To get from cipher text to plain text, you just subtract. Here’s a graphic of a Vigenere square to help you visualize this.
Now, you can do the same thing I’ve outlined above without a repeating keyword. This is what’s done with so-called “one-time pads” or with ciphers that use long strings of characters (such as books or other texts) to encode the message.
Now, if you’re being sneaky, you can add all sorts of complexity to this. For example, you could throw in numbers at the beginning or end of alphabet to add extra characters. For example, if instead of using an A to Z character set, your character set goes A-Z,0-9. In that case, using my example above, “I+P” would become “1” instead of “B”.
Just throwing out some ideas…
Lemme see if I’ve got this straight:
ABCDEFG - Keyword
SECRETZ - Plain text
-----------------------
TGFVJZ7 - Encrypted text
So A, being the 1st letter of the alphabet, is equal to 1. It would be added to S, the 19th letter of the alphabet, giving one a sum of 20. Since T is the 20th letter, it would be the first letter of the encrypted code. B(2) plus E(5) would equal G(7). And so on?
A is equal to zero. B is equal to 1… Z is 25.
I downloaded Huckleberry Finn from the Gutenberg Project, removed the preface and epilogue and numbered the paragraphs. I included the chapter headings because I don’t know how to remove them automatically, and I wasn’t about to go through the entire book chapter-by-chapter. (Sorry.) Anyway, here is is: http://bit.ly/QY9gr. I’ll leave it to the other wizards on the board to figure out what the heck to do with it.
Going off of the “hint” that you got, as far as I can tell, the first word might have a double vowel in it. This would make it easier to figure out if it is a Vigenere cipher. I’m trying to find a keyword that also has a double vowel, so the first letter and the next two line up, but so far I haven’t come up with anything. There’s just too many variables in what might count as paragraphs.
Of course, this could be a complete dead end, and we’re barking up the wrong tree.
What’s your hunch on how the numbers figure into the cipher? I have a feeling it may not be a straight Vigenere, but more like a polyalphabetic cipher encrypted with the letters starting at the denoted paragraph. I’ll try looking into a couple of different key lengths using a Vigenere, but, my hunch of a keylength of 4 (based on the spacing of a couple of digraphs and trigraphs), does not work out, as the frequency of letters and numbers is scattered all the way from 0-9, A-Z. To be clearer, assuming the decoded text is a normal shifted Vigenere using either a sequence of {0-9,A-Z} or {A-Z, 0-9}, one would expect a gap of frequencies for where the numbers 0-9 should align (since the Vigenere is just a shift in letters, not a scrambling of them). For a key length of four, there is no such sufficient gap. I will try other key lengths later to see if I have any more luck.
I’m pretty new at this, so, my hunch is of pretty low value. 
I’m guessing he’s mapping the numbers at the end, but who knows. If this is supposed to be a hint, it’s almost as difficult as the original code, so I’m hoping he’s built in some easy to spot clues like the “VXXC”. Since the shift works out along a chain, I’m hoping we can work out that first word, and most everything will fit in place. But really, I have no idea. Those numbers are still bugging me. I feel there’s no obvious starting point here, which is sort of disheartening. This might be well over my head (which is no big surprise).
I twittered him this message: Vigenere cipher? 56th paragraph from the book Huck Finn? If so, start from Introduction? Chapter Markers count? Or, wrong path?
I hope he responds with something helpful…
I’ve tried key lengths from 2 to 12, and none of the frequency distributions give me anything that looks promising. Now, it’s possible there are just a lot of numbers in the coded text. If that’s the case, that might be throwing my distributions off. I just don’t see any pattern in the frequency distributions that match the general frequency distribution of the English alphabet. My gut is still telling me it’s not Vigenere.
Agreed. The distribution of the numbers makes no sense. I’ve got nuthin’.
It’s really starting to get me wondering what is going on with this code. I love ciphers, but I can’t quite place this one.
I have a breakthrough! Be back in a few minutes! Woo hoo!
The message on the twitter page reads:
YOU GOT THIS FAR BUT THERES A DIMENSION TO THIS PUZZLE THAT YOU HAVE OVERLOOKED XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
So, it is not a Vigenere, but a polyalphabetic cipher with a very long key and some additional complexities, and Paragraph 2090 corresponds to the paragraph 2090 in the project Gutenburg edition of Huck Finn. The text file Here Comes Dots posted contains paragraph numbering, and Chapter headings are marked as paragraphs. As a hunch, I tried shifting both forwards and backwards. Turns out the backwards shift was correct. And the scheme of character mapping was ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890. (I initially had “0” between ‘Z’ and ‘1’, but that worked itself out quite quickly.)
So, we have the following (ciphertext, followed by key, followed by decoded text):
iwz3tbi5xajt1yvcu1fgn7osu2sux1pu992v31ydc9zzdmu9hxd2fdz9fj0nprvsw6300h300mcigc63fg304g6g630cc9yhh6yhy7ff
theverywordsiwasasayinnolongeragothnthisminutetosisterutterbacknshelltellyousoherselfshshelookatthatairr
YOUGOTTHISFARBUTTHERESADIMENSIONTOTHISPUZZLETHATYOUHAVEOVERLOOKEDXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
So, in this code, the letter “a” means go back one position in out character map, “b” means 2, “c” is 3, and so on. Now, to find the missing dimension to the puzzle.
The type of cipher is known as a running key cipher, for those interested.
Wow! Great work pulykamell!
So, what have we overlooked…?
I wonder if we missed something on the initial Fibonacci puzzle?
There is an awful lot of so-far-unused information in the original puzzle. It would be so much more elegant if the extra information was usable. However, it would be elegant only if information from the second puzzle was required in order to get the information out of the spiral.
A coupkle othe musings:
- “Dimension” is a very particular word. I note that, including the paragraph number, there are 27 (3X3X3) groups of four in the current puzzle.
- Seems like this persons Twitter name (threeoutoffour) ought to come into play at some point.
Well, here’s my observations:
So far the puzzles have been fair. By that, I mean there hasn’t been any intentional misdirection/red herrings. What strikes me about this latest one is the string of 39 Xs that end the cipher. I get the sense there must be a clue in that. There’s no reason for the message to end with that many characters. The fact that they’re all Xs is there, I believe, in part to show that we’ve done our deciphering properly. In other words, if there was simply a random strip of 39 junk characters at the end, a logical assumption would be that the cipher somehow changes at the end. The puzzle writer made sure, by giving us Xs all the way through to the end, that we know we did everything correctly, and there’s no additional message to be found in the ciphertext itself at the end, so we’re not wasting time trying to decode the end of the message.
On the other hand, why 39 Xs? The puzzle writer could have simply ended the message there, with no extra Xs. This leads me to believe (and I’m fairly confident) that the final string of Xs is a clue to the “overlooked” part.
I like the idea that it’s a clue with what to do with some of the leftover characters in the Fibonacci puzzle. I’m not married to that idea, but it’s a reasonable guess, and would make that puzzle more elegant. Still, if it’s not, I won’t feel “cheated” by the Fibonacci cipher. A skip cipher is a perfectly reasonable and relatively well-known method of encoding. The junk letters don’t have to mean anything, but it would be cool if they do.
The choice of the word “dimension” also pinged my Spidey sense. I’m not sure what to make of it yet, but it’s possible that’s a clue, too, as suggested by zut. It’s a common enough word that it may just be the author’s natural choice of words, but it’s just odd enough that it’s worth considering as a clue to solving the puzzle.