Re: Some entropy calculations with secret/huffman interaction

The paper I linked to some time ago has one analysis of this.
-=R


On Wed, Nov 27, 2013 at 11:00 AM, Ilari Liusvaara <
ilari.liusvaara@elisanet.fi> wrote:

> On Wed, Nov 27, 2013 at 09:08:04AM -0800, Martin Thomson wrote:
> > On 27 November 2013 02:40, Ilari Liusvaara <ilari.liusvaara@elisanet.fi>
> wrote:
> > > The secret
> >
> > Have we worked out yet how to determine if something is secret or not
> > algorithmically?
>
> No, I just wanted to calculate how bad the entropy loss is.
>
> Also, doing Huffman on anything that looks like base64 probably isn't a
> good
> idea.
>
> Hexadecimal stuff (escpecially with lowercase a-f) does compress
> (average ~30%; Hmm... I wonder how badly those leak...).
>
> -Ilari
>
>