Rsync mirroring interacting with compression

[Michael James]

Soon I'll be maintaining a mirror of a large collection of datasets.
They are big and change slowly over time.
They come as a gz compressed ascii file.

I know I could just script downloading (FTP) and unpacking them.

Or maybe the server allows an rsync download.
So I maintain a mirror of the source directory and rsync it each night.
Then script the unpacking from there.

But how does a gz file react to changes in the source.
Is it likely to change large sections of the compressed file
 for a small addition to the source, like a hash?

Is there a more rsync-friendly compression available?

Thanks for any insights,
michaelj

-- 
Michael James				michael at james.st
Network Programmer			voice:	02 6246 5040