|
May 10, 2005
So, what do we do if/when frequency analysis gives us the wrong
answer. This is a very likely outcome as letter frequencies in encrypted
text is unlikely to match any "standard" character mapping exactly.
Here are some ideas:
- Multiple FreqLists - it may be useful to apply more than
one FreqList when trying to decrypt a file. Different text sources
(like "literature" versus "the dictionary" versus "speeches", etc)
will have different frequency characteristics, and you may get
better results from one FreqList over another.
- Interactive - you can create a little interactive
software tool to change a CharMap used for decryption and then show
the effect on a couple of sentences... sort of a trial and error
thing.
- Semi-exhaustive - so, okay, we can't try each of the 26!
possible mappings (4 x 10^26 or roughly half an
octillion),
but maybe you can pare the problem down to just the top 10 letters
or so. Watch out though... factorial grows fast:
| 1! =
1 |
2! = 2 |
3! = 6 |
4! = 24 |
| 5! =
120 |
6! = 720 |
7! = 5,040 |
8! = 40,320 |
|
9! = 362,880 |
10! = 3,624,800 |
|
11! = 39,916,800 |
12! = 479,001,600 |
- Common words - it may be useful to dig up a list or two
of common words. Try a number of CharMaps and keep the one that
generates the greatest number of "common" three letter or four
letter words. I think there's a link to common words on the main
Program #4 page.
- Crytopop? - here's this guy Crytopop who has some thoughts on using letter frequency
for decryption:
http://www.cahlander.com/crypto/cryptopop.html
Just some food for thought.
yow, bill
|