Cryptography Reference
In-Depth Information
∼
0.87 % 'ˆD'
0.86 % ' '
0.17 % '
'
0.17 % '
<
253
>
'
0.84 % 'ˆP'
0.17 % '
<
251
>
'
0.83 % 'ˆC'
0.17 % '
<
234
>
'
0.80 % 'ˆF'
0.17 % '
<
189
>
'
0.78 % 'ˆB'
0.17 % '
<
183
>
'
0.78 % '4'
0.15 % '
<
249
>
'
0.72 % 'ˆ@'
0.14 % '
<
239
>
'
...
0.14 % '
<
236
>
'
0.18 % '
<
250
>
'
0.14 % '
<
219
>
'
0.18 % '
<
223
>
'
0.11 % '
<
174
>
'
0.18 % '
<
221
>
'
0.10 % '
<
255
>
'
Figure 3.14:
Character frequencies in a compressed file.
opinion), though it does make cryptanalysis a bit harder. The illusion of more
security comes from the fact that the character distribution curve is very flat
in compressed files; there are almost no usable maximums and minimums (see
Figure 3.14).
In contrast to natural language, however, compressed text obeys fixed rules —
something we'll take full advantage of. We try to mount a ciphertext attack
against Vigenere-encrypted compressed files, referring to
compress
, a program
that once was widely used in the UNIX world.
But first, we have to deal a bit with the plaintext or, more specifically, with
the format of files created by
compress
.
How Are Files Compressed?
Most people realize that language is not the shortest form in which we can
express ourselves. Compression means that a piece of text is 'summarized'
in a way that ensures no information is lost. More specifically, a program is
