Algorithm that makes maximum compression of completly diffused data.

[jonas.thornvall at gmail.com]

Den lördagen den 2:e november 2013 kl. 22:31:09 UTC+1 skrev Tim Roberts:
> jonas.thornvall at gmail.com wrote:
> 
> >
> 
> >Well then i have news for you.
> 
> 
> 
> Well, then, why don't you share it?
> 
> 
> 
> Let me try to get you to understand WHY what you say is impossible.  Let's
> 
> say you do have a function f(x) that can produce a compressed output y for
> 
> any given x, such that y is always smaller than x.  If that were true, then
> 
> I could call f() recursively:
> 
>     f(f(...f(f(f(f(f(x)))))...))
> 
> and eventually the result get down to a single bit.  I hope it is clear
> 
> that there's no way to restore a single bit back into different source
> 
> texts.
> 
> 
> 
> Here's another way to look at it.  If f(x) is smaller than x for every x,
> 
> that means there MUST me multiple values of x that produce the same f(x).
> 
> Do you see?  If x is three bits and f(x) is two bits, that means there are
> 
> 8 possible values for x but only 4 values for f(x).  So, given an f(x), you
> 
> cannot tell which value of x it came from.  You have lost information.
> 
> -- 
> 
> Tim Roberts, timr at probo.com
> 
> Providenza & Boekelheide, Inc.

Well let me try to explain why it is working and i have implemented one.
I only need to refresh my memory it was almost 15 years ago.
This is not the solution but this is why it is working.
65536=256^2=16^4=***4^8***=2^16

Yes i am aware that 256 is a single byte 8 bits, but the approach is valid anyway.