Re: Need help solving a puzzle. Please.

From: William L. Bahn (william_at_toomuchspam.net)
Date: 09/06/05 

Date: Tue, 6 Sep 2005 00:46:59 -0600

"Kristian Gjøsteen" <kristiag+news@item.ntnu.no> wrote in message
news:dfi871$uuo$1@orkan.itea.ntnu.no...
> daytribe <mike.day@techie.com> wrote:
> >Do I have to work down the columns or across the columns?
>
> Read it like you would normal text. And ignore the five-letter
> blocks. Traditionally, ciphertexts are grouped into five-letter
> blocks without any regard for the underlying cryptography.
>
> >Do the numbers refer directly to the position of letters?
> >
> >Are the numbers in pairs for row and column e.g. 29 is row 2,
column 9?
>
> Try counting the number of words in the key. I found that
pretty
> suggestive.

I just came upon this thread and am having trouble as well - not
surprising since I am VERY new to this game.

Given the puzzle statement, I can see that it is likely a book
cipher with the passage of text at the bottom being the key (and
this makes sense given the "key lay in the passage beneath"
phrase in the puzzle statement.

But the question then is what is the algorithm used to extract
the plaintext from the key text? The total number of digits in
the ciphertext is 273. Since this is not even, that would suggest
that the ciphertext does not consist of column/row pairs. The
count factors as 3x91 and, since 91 is prime, would suggest that
the ciphertext should be broken into groupings of three. Of
course, I am assuming that there aren't any chaff digits intended
specifically to fool this type of analysis. If column/row pairs
are used, they could conceivably be encoded as three digit
groups. Are these reasonable suppositions to make, at least
initially?

The number of words in the key is 96. I don't find that
particularly suggestive beyond being close to 100 which would
mean that the words could be organized into a nearly complete
10x10 grid, but then we are back to the row/col pair issue.

There are 480 alphabetic characters in the passage.

Breaking the ciphertext into three digit groups, I find that if I
assume that each group is the character position from the start
of the passage, that I am immediately in trouble because there
are numerous entries greater than 480. If I assume that it is a
row/column pair then the most obvious choices would be RCC or
CCR. Since there are 11 rows in the key text (assuming that the
line breaks in the original post are the ones that should be
used), then RCC yields column positions as high as 97. The
longest line only fewer than 60 characters in it (only a bit more
if the spaces and punctuation are left in). The CCR choice does
me no better, as it also has values in excess of 90. Of course,
it could be simply that the count wraps around.

If I take a different tack altogether, I might assume that each
digit in the ciphertext is an character count from the previous
digit. Assuming only alphabetic characters are to be used, then
the first two groups (29150 31235), might decode to "HIDEE ISSNT"

29 15 03 12 35 25 12 23 61 12 41 72 1

Since this means working with either row zero or column zero, I
will assume base-zero indexing.

RC: 29 doesn't exist since Row 2 (third row) only has 9 words in
it (need ten).
CR: and shelter soil of many for of wide slopes of the or

Taking the words doesn't make any sense. Taking the first
character of each word leaves me with "ASSOMFOWSOTO", which
doesn't look much better.

Looking at the 96 words available, the first letters do not yield
a very rich collection of characters which makes me think that
the letters within words are also fair game. That might make the
groups of three make sense with one digit being a row, one being
a column, and the third being the offset within a word. Looking
at the initial characters, fifteen of the ninety-six words begin
with the letter 'O' and just a few other letters account for a
large fraction of the rest. Although, somewhat to my surprise,
there were only seven letters ( HJKLXYZ) that never appear as
initial characters of words, so perhaps I am barking up the wrong
tree.

I'm guessing that I am making this much, much harder than it
needs to be, since you indicated that your very first attempt to
decrypt characters led to sensible results.

What kinds of things should I be looking for?

Relevant Pages

  • UK postcode regular expression (was: String comparing)
    ... > always being single digit and two letters. ... characters long; the outward code (the ... part before the space) is two to four characters long and the inward ...
    (comp.lang.php)
  • Re: Problem opening a web page in firefox that has a hyphen "-" character.
    ... "Note that while upper and lower case letters are allowed in domain ... start with a letter, end with a letter or digit, and have as interior ... characters only letters, digits, and hyphen. ... subject of leading hyphens in URLs, but perhaps I used the wrong search ...
    (Ubuntu)
  • Re: How do I format cells to a specific number of digits?
    ... this will then stop numbers with 7 digit from begin entered. ... > move the mouse over the bottom right hand corner of cell B1 and your mouse ... The same happened with letters. ... How can it be formatted so only 2 characters are ...
    (microsoft.public.excel.misc)
  • Re: User Accounts
    ... Change the name of the account. ... hackers as a means of getting a foothold into your system. ... using all upper case or all lower case letters. ... It should contain at least eight characters. ...
    (microsoft.public.windowsxp.help_and_support)
  • Re: Postroad Mailer message file name algorithm
    ... possible filename characters, 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ. ... The first digit of a *.POP filename is the last digit of the current ... into that array of characters. ...
    (de.comp.os.os2.networking)