Kiddie Codes

[HaB]

So I have this younger cousin. He's really into fooling around with codes/ciphers and such. I was into the same stuff as a kid, so I figure I'd save him all the work of encoding and decoding by hand, and write him a coupla cool little programs that would do it for him.

So there's a code table I'd like to generate for him, but I'm not sure of the best way to store the data for easy retrieval later. The table works like this:

• You take a key word, having as many non-repeating letters as possible. We'll use 'exchampion' for our example.
• You write out that keyword, followed by the remaining letters in the alphabet in order, using columns of 10 letters, like so:

       -------------------
       e x c h a m p i o n
       b d f g j k l q r s
       t u v w y z . ! ? ,

       The last 4 chars are just here to complete the table.

• Then, rearrange the columns in alpha order according to the topmost letter, and number them:

    1 2 3 4 5 6 7 8 9 0
    -------------------
    a c e h i m n o p x
    j f b g q k s r l d
    y w t w ! z , ? . u

• Finally, number the rows from 1 to 9 vertically, 3 numbers on each row, like so:

          1 2 3 4 5 6 7 8 9 0
        |---------------------
  1,4,7 | a c e h i m n o p x
  2,5,8 | j f b g q k s r l d
  3,6,9 | y w t w ! z , ? . u

Now our table is complete. The way to use it is...you find the plaintext letter you want to encrypt, the ciphertext then becomes a two number set made of a randomly selected row number for that row, and the column number. So, plaintext 'a' becomes ciphertext 11, 41 or 71, 'b' becomes 23, 52, or 83, etc, etc.

I can figure out ways to generate the table, given a user supplied keyword, but I'm wondering what the best way to store it so it can easily be looked up against? My initial impression is to use 2 hashes, one with the plaintext as keys, the values of which are an array containing each of the 3 possible ciphertext values, and the other being the opposite, but is there a more elegant solution? Some way to store a matrix of data like that that I'm not thinking of?



Thanks!

HaB

[tedv]

Just for fun, I wrote this script to parse /usr/dict/words looking for long words without repeating letters. Here's the code:

#!/usr/bin/perl -w
use strict;

my @words = ();
my $max_len = 0;
my $delta = 2; # Number of letters less than maximum which is acceptab
+le

while (<>) {
    # Check for words which are too short
    my $cur_len = length;
    next if $cur_len + $delta < $max_len;

    # Check for words with repeating letters
    next if /(.).*\1/;

    # Store value in list, possibly destroying old values
    if ($cur_len > $max_len) {
        $max_len = $cur_len; 
        @words = ($_);
    } else {
        push @words, $_;
    }
}

print @words;

And if you're curious, here's the output:

ambidextrously
bankruptcies
bluestocking
configurable
considerably
consumptively
copyrightable
customizable
exclusionary
flowcharting
incomputable
productively
questionably
recognizably
unpredictably
unprofitable
upholstering

A thirteen letter word might work better because you wouldn't have to pad the cypher.

-Ted

[Fastolfe]

I might do something like this:

$letter = 'a'; @ct = qw{ 11 41 71 };
$forward{$letter} = [ @ct ];
...


# Build reverse hash
foreach $letter (keys %forward) {
   foreach $ciphertext (@{$forward{$letter}}) {
      $reverse{$ciphertext} = $letter;
   }
}

print join(" ", @forward{split(//, "plaintext")});
print join("", @reverse{qw{ 19 59 71 75 77 63 13 66 33 }});

[TGI]

For the cyphertest lookup try using the indexes of your arrays as keys. This should be speedy and cheaper from a memory standpoint.

Split your cyphertext 'letters' into two digits. Then create a list of list references.Something like this:

@ct[2,5,8]=['d', 'j', 'f', 'b', 'g', 'q', 'k', 's', 'r', 'l'];
print $ct[2][3]; # Should print b.

Warning:code untested as all get out.

[HaB]

Minor correction: In the example above, a plaintext 'b' would become ciphertext 23, 53 or 83, NOT 52. My apologies.

-HaB