Date   : Mon, 18 Sep 2006 07:50:20 +0100
From   : Sprow <info@...>
Subject: Re: Compressed ROMFS?

In article <4e67ac7d89info@...>,
   Sprow <info@...> wrote:
> In article <20060916095434.53696.qmail@...>,
>    Greg Cook <debounce@...> wrote:
> > Did anyone ever come up with a system for compressing files in ROMFS
> > images?  I sketched out a dictionary based algorithm yesterday but the
> > hard part is finding the repeating substrings.
>
> I had a quick bash this morning at Huffman coding some data
>
> For tokenised BASIC you can save 21%
> For 6502 binaries you can save 10%

As a postscript to my own message, I tried some bitmapped data too (very
high proportion of 0's for black) and that compressed by 61%.

Lastly I wrote an RLE compressor using the madness proposed in
<4e5b03df85info@...> and that gave
  For 6502 binaries you can save just 1%
  For tokenised BASIC you save -2% (ie.bigger!)
  For bitmap images with high black content 41%
Sprow.