Article · October 008 citations reads 5,930 authors

48
Dr. Dobb’s Journal
l
www.ddj.com
l
October 2008
Core Technology
THE BOOK CIPHER ALGORITHM
guessing probable plaintexts along the
ciphertext, attackers will eventually recog-
nize the book and break the code.
A better idea is to replace words in the
plaintext with the location of words from a
book. For example, to encode the word
“computer,” you look for the first appear-
ance of “computer” in the previously cho-
sen book and enter its position as the
cipher text. The second appearance of the
word “computer” is replaced by the posi-
tion of the second appearance of that word
in the book, and so on. The real problem
with this approach is finding the word: If
you agree to use David Copperfield as a
code book, and then try to encrypt an arti-
cle about hash functions, you are unlikely
to find all the necessary words. 
An alternative approach that gets around
this problem is to replace individual letters
rather than words, in which case the Book
cipher is properly a cipher. Figure 2 illustrates
the concept. We are encoding a message
“DRDOBBS” using the same passage from
Harry Potter and the Order of the Phoenix.
To code the letter “D,” we look for the first
word in the passage starting with “D” (it’s the
6th word, “did”). Then we look for the first
word starting with “R” (the 11th word,
“rang”), then for the next word starting with
“D” (the 16th word, “down”), and so on. The
final ciphertext is 6, 11, 16, 17, 2, 10, 15.
Decoding is even simpler. We take the
ciphertext number-by-number and look for
the corresponding words in the book, gener-
ating the original plaintext.
Practical 
Issues
The Book cipher is straightforward to imple-
ment and secure, but using it in real life pres-
ents some difficulties. First of all, you can
encrypt only letters, preferably after convert-
ing them to uppercase. Most of the crypto
textbooks deal with uppercase letters only—
introducing symbols that appear very often
(such as blanks) or sparsely (such as $ or the #)
gives attackers an unnecessary advantage.
On the other hand, sometimes you have to
encode a message containing numbers or
special symbols, and reading decoded text
without blanks is a real pain.
This issue can be resolved by pre-
encoding the plaintext using an algorithm sim-
ilar to UUENCODE, which produces upper-
case letters only. A possible candidate is the
Five-Letter Codegroup Filter (www.fourmilab
.ch/codegroup). While encoding, you could
compress text and even implement some
kind of double encryption. Compression is
continued from page 46