Re: 3.5" Disk archiving

[Jeff Blakeney <CUTjblakeney@sentex.net>]

On Mon, 17 Feb 2003 02:59:15 GMT, Andy McFadden <fadden@fadden.com>
wrote:

>You are correct, but I think you've reached the point of diminishing
>returns.  ShrinkIt does an RLE pass before feeding the data into LZW,
>so an empty 512-byte block is going to go down to nothing before LZW
>even gets near it.  You're not going to lose a lot of ground by having
>the RLE bytes in the LZW trie.

Ah.  I didn't realize that ShrinkIt did RLE encoding before
compressing.  Well, I might have but it has been such a long time
since I've looked over the format that I don't remember.  :-)  It
would definitely give better results if the RLE was done on the entire
contents of the disk rather than just each block.

While I'm dreaming, it would also be nice if the archiving program
were able to check to see which blocks are actually in use and only
pack them and then rebuild the disk when it unpacks it.  Maybe with
Unisys' patent expiring this summer we can finally get some updates to
ShrinkIt and GSHK.

>Of course, you could also argue that you will get better compression if
>text file sectors aren't intermixed with binaries.  Copying the files to a
>new disk should yield sequential allocation, which probably provides more
>of a compression boost than clumping the zeroed blocks together.  For best
>results you should copy different categories of files in different groups,
>kind of like "solid" archiving in RAR.

That is true and is something I hadn't really considered.

However, to make things clear, the point I was trying to get across is
that disk archives don't have to be as bad as some people make them
out to be.

-- 
 Jeff Blakeney - Dean of the Apple II University in the
                 Apple II Community on Syndicomm.com
 CUT the obvious from my address if you want to e-mail me