Re: Brute Force (+n cipher)

[Brian Eric Farnell <bfarnell@xxxxxxx>]

Claus Anders wrote:
> > where I simply replace the nth char. in the word with same letter+n
> > in the alphabet (with an one to one char mapping).
> 
Gabriel Landini wrote:
> I wonder what the character distribution and the entropy of this
> would be. Each word is always written the same, but the word
> construction is completely different.

Just a guess off the top of my head, the effect on entropy could
be very different depending on the language.  For example, where
two letters often occur together at the beginning of the word,
they will still always be reflected in an analogous pair,
however, if you take a language heavy in endings after an
variable number of letters like Russian, those values get
muddled.  Basically, all contributions to predictability from
initial sequences and/or prefixes stand as they are.  I'm not
sure what the splitting of each standard ending/suffix into 4 or
5 analogous pairs would do mathematically, my brain keeps going
one way then the other on it.
Regards,
Brian