🚀 go-pugleaf

In article <1992Nov18.134243.24089@qiclab.scn.rain.com> Leonard.Erickson@f51.n105.z1.fidonet.org writes:
>jdailey@dadd.ti.com (Jim Dailey) writes:
>
>>So what are some of the methods used to attack an encrypted text, when the
>>encryption method is unknown?
>
>First you run frequency tests. This will tell you *immediately*
>if a "reasonably sized" text was encrypted via *any* character
>transposition cipher. 

This thread is at about my speed - so I'll toss in a question.  I'm sure
that this is hopelessly naive - but I hope someone will tolerate the
question long enough to answer it.

Suppose that I have a character transposition cipher.  Will the
following scheme make it appreciably stronger (or weaker?):

To encrypt:

    0) compose the message
    1) add (to both the front, and the back) additional characters,
       chosen to produce a flat histogram.  Pseudo-randomize, as needed.
       Perhaps also pad to a standard size block?
       Perhaps randomize the positioning of the message within the block? 
       Perhaps break the original message into bite-sized pieces and
       include additional material between bites as well as at the
       beginning and end?
    2) apply the character transposition cipher.

to decrypt:

    0) reverse the transposition
    1) rely on the reader (either human or otherwise) to recognize
       the start and end of the original message (or the multiple starts
       and ends of message bites).

One weakness which springs to mind is that absolutely flattening the
character histogram may cause an unacceptable increase in the size of
the transmitted message.  Perhaps this is what was intended by the
qualification ``"reasonably sized"'' above?

A (probably foolish) complication which springs to mind is to:

    *restrict words in the message to a pre-defined lexicon
    *break up the original message on word boundaries
    *add "histogram-flattening" characters between words (instead of
     whitespace?) 
    *ensure that no "allowed words" are generated accidently
    *have the decrypting process find the "allowed words" among
     the random garbage (after reversing the transposition)

Assuming that the message has been appropriately un-transposed, it
should be simple enough for the intended receiver to find the allowed
words, and hence the original message.  The question is: does this give
the enemy too large a lever on cracking the transposition cipher?





-- 
Kenneth Sloan                   Computer and Information Sciences
sloan@cis.uab.edu               University of Alabama at Birmingham
(205) 934-2213                  115A Campbell Hall, UAB Station 
(205) 934-5473 FAX              Birmingham, AL 35294-1170