Re: VMs: Reducing the 2nd order entropy by sorting within words

[Gabriel Landini <G.Landini@xxxxxxxxxx>]

On Thursday 07 Aug 2003 10:12 am, Anders, Claus wrote:
> Just for fun I wrote a small awk program and let that work with some Latin
> text.It just sorts the characters with a word unique and guess what: the
> output h2 is as low as 2.277 (according to monkey).
[...]
> Seems, that sorting could be part of the encryption scheme.

I remember Rene's letter sorting as well, and I always wondered whether the 
resulting text could be reconstructed. The dictionary idea would help. The 
problem is that there are many solutions to a single character sorted word:

eehrt -> three, there ...

Character sorting would mean that one could not see word patterns in which the 
letters go "backwards" (like "abca" and so on, since the sorting would put 
the two "a" together).

Below are the frequencies of the most common 22 patterns (I found 969 word 
patterns). The arrow "<---" shows those patterns that are not "increasing".

freq pattern
---- -------
5716 abcde
5610 abcd
3465 abc
3188 abcdef
2500 ab
2307 abccd
1548 abcdde
1282 a
1166 abcdefg
804 abcdeef
778 abbc
640 abccde
516 abcddef
448 abac <---
347 abcdefgh
335 abcbd <---
303 abcdeffg
290 abcad <---
274 abcdbe <---
263 abcda <---
245 abcdae <---
234 abacbd <---

cheers,

Gabriel

______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list