Subj : Numbers stations
To   : Angus McLeod
From : Deuce
Date : Thu Aug 18 2005 02:35 am

  Re: Numbers stations
  By: Angus McLeod to Deuce on Wed Aug 17 2005 23:00:00

 >  > Let's say I'm using a current magazine as the OTP and the numbers represe
 >  > page, line, word, letter.  I wouldn't need to repeat the same set of four
 >  > numbers at all ever.  Without any repetition, there is no pattern to disc
 > 
 > That isn't exactly how an OTP works.  You are describing a perectly
 > workable method of encoding data (a la 'The key to Rebecca') but it really
 > doesn't qualify as am OTP.  Not as far as I understand it.
 > 
 > Besides the data for an OTP should be as random as possible.  The text of
 > a magazine would never be random enough to be considered for a page of an
 > OTP.

The thing is that the randomness derives from both the text and the choice of
the letters from the text.  That is to say rather than using a pad and
encrypted text of the same length, you are randomly choosing the pad from a
much larger possible pad.

However, it's poor use of a pad for the basic reasons that the pad is available
to anyone who cares to dig it out, and the pad is not destroyed after use (Both
stem from the fact that more than two copies of the pad exist).

To properly encode something using this method, the choice of letters should be
as random as possible... though I can't think of a way of squeezing more than
about 50% entropy into the data. Lucily, it's not the randomness per se that's
the ssue, it's the predictability.

---
 þ Synchronet þ ``Penguins make tasty snacks''

.