(SPOILER EDITION) Descent of the Durtle into Madness


#141

But how do you get from 3 to E and 5 to G? And sometimes you shift in different directions.

Also O is not 0.


#142

well… 3 looks like an E :grimacing: and a 5 like a G, also 5 is ご

53NKI I

And I shift right or left… whatever I feel like. I know… it’s too farfetched


#143

Letters for numbers

  • 0 = O or D or space
  • 1 = I or L
  • 2 = Z or e
  • 3 = E or :heart:
  • 4 = h or A
  • 5 = S
  • 6 = b or G
  • 7 = T or j
  • 8 = X
  • 9 = g or J

Even worse :stuck_out_tongue:


#144

I’m with @emucat in still trying to decode the previous codes just to try and verify if they are just a means to get to the other pages we found with comments or determine if they are still part of the puzzle.

I found a pattern that partially links up between the codes and the pages found, but it doesn’t 100% fit.

Recap of the pattern:
AABQD010L76223CEGD72FI5DPBOJJMO+
B51BIQPGS206HD
00LB269G
D02315CPQAWDF0B~IL7KB1FG+A

I decided to count the number of characters in each encrypted row and you get:
32
14
8
26

I then cross-referenced with the Tofugu articles that were discovered and found a similarity between the codes and the URL of each article. Now remember, below our four encrypted clues it said “Caution! Space is also required.”, so each hyphen below (used in place of spaces in the URL) is included with the count:
32 how-to-install-japanese-keyboard
14 genki-textbook
8
26 beginner-japanese-textbook

So 3 of the 4 pages with comments fit the pattern, but the page that does not fit has the comment “Durt Durt! Nice try!” which makes me think it was a false trail we may have found through some other means.
The page that does not fit is:
36 dictionary-of-basic-japanese-grammar

And we don’t have anything matching 8 characters. So maybe there is one article out there we are still missing.

Thoughts?


#145

image
Coincidentally I was looking at it the same way, but didn’t look further once I came across the first length mismatch.

If you want to pull that thread some more, here’s all the 22 article URL’s where the last segment in the path is eight characters in length:

Summary

https://www.tofugu.com/travel/hanachan/
https://www.tofugu.com/japanese/otonagai/
https://www.tofugu.com/japan/washitsu/
https://www.tofugu.com/reviews/tonoharu/
https://www.tofugu.com/japanese/tsundere/
https://www.tofugu.com/japanese/tsundoku/
https://www.tofugu.com/japan/ashigaru/
https://www.tofugu.com/japan/butsudan/
https://www.tofugu.com/japan/japonism/
https://www.tofugu.com/reviews/hinative/
https://www.tofugu.com/japan/namahage/
https://www.tofugu.com/japan/mimikaki/
https://www.tofugu.com/interviews/hinative/
https://www.tofugu.com/japanese/gyaru-go/
https://www.tofugu.com/japan/pachinko/
https://www.tofugu.com/japan/purikura/
https://www.tofugu.com/interviews/altjapan/
https://www.tofugu.com/japan/nurikabe/
https://www.tofugu.com/japan/bosozoku/
https://www.tofugu.com/japan/botaoshi/
https://www.tofugu.com/reviews/skritter/
https://www.tofugu.com/japanese/furigana/


#146

Maybe the furigana one? Pretty sure none of the others match up with the theme.


#147

This doesn’t fill me with joy, considering that’s the page we found the codes on in the first place.

Question is, is there any way to reverse-engineer a black box algorithm that renders “how-to-install-japanese-keyboard” as “AABQD010L76223CEGD72FI5DPBOJJMO+” and then see if it works with the other codes too?

“hinative”, “skritter” and “tonoharu” are all reviews. But yeah, not reviews of the books we’re looking at.


#148

Working on analyzing that text. Considering rotating keys are used to encode this cipher…


#149

That looks good!


#150

So for a simple XOR cipher, you can recover the key if you have both plaintext and ciphertext by xor-ing them back together. If it’s that, AND if the key is the same for all 4 items, we can get the unknown ones.

K = C1 xor P1
then,
Pn = K xor Cn

complication: how the xor works depends (once again) on the encoding scheme of the characters. ASCII? Unicode? etc.


#151

What are we XORing, though? ASCII values?


#152

That’s not a complication because the Latin range of both schemes is the same. Naturally, that ensures backwards compatibility of documents encoded in ASCII when they are re-encoded in Unicode. Or put more specifically, Captial A is 0x41 and U+0041 in ASCII and Unicode, respectively.


#153

Still, I have done a lot trying to XOR against this text before we discovered the book cipher elements, to no avail. That’s why I’m exploring Vigenere most recently. In part based on a hunch that if the puzzle IS NOT using cryptographic techniques that require computer science like XOR, then the puzzle is friendlier to a larger user base.

EDIT: But I like XOR because it gives a more plausible explanation for the appearance of the symbols in the ciphertext like ~ and +. Sigh so much to speculate

EDIT: If someone likes to play, this site has lot’s of tools to attack https://www.dcode.fr/vigenere-cipher

I used this one more for the XOR, it’s SOOO stupidly good versatile, but a less friendly to non techies. This link will load it with the recipe that tests the entropy of the ciphertext. Don’t be misled by the scale when it shows that this text doesn’t look like properly encrypted text. Take that to mean, if it’s encrypted it’s “weak” as in not a proper industry-grade level of encryption. https://gchq.github.io/CyberChef/#recipe=Entropy()&input=QUFCUUQwMTBMNzYyMjNDRUdENzJGSTVEUEJPSkpNTytCNTFCSVFQR1MyMDZIRDAwTEIyNjlHRDAyMzE1Q1BRQVdERjBCfklMN0tCMUZHK0EK


#154

… How did I not see that?


#155

Well, I went through the source code for all the likely ones and saw nothing, so hmm.

I’ll keep looking but I don’t think this is it.


#156

How are we doing? Do we have more words? Leads?


#157

I’m personaly working on the Adventures in Japanese page but I’m not finding anything.


#158

I tried that one and I got:
eymtof


#159

Aye, that’s what I got too. Or “tamscl” if I include the heading.


#160

we should add it to the top post so nobody else tries