Yes, that’s easy:
dd if=/dev/zero count=1000 | gzip --verbose --stdout | \
gzip --verbose --stdout > /dev/null
1000+0 records in
1000+0 records out
512000 bytes (512 kB) copied, 0.00438917 s, 117 MB/s
99.9%
93.4%
But you should be asking whether *useful* recursive compression has ever
been demonstrated.
>
> So, I thought to ask here if anyone has actually done such.
>
>
> This topic has been the domain of claims and no proof as far as I know.
>
Many, many claims, but also plenty of proof that it couldn’t work in any
real useful sense.
>
> I have a scheme to try in the near future myself.
>
Well, recursive, infinite (i.e. down to zero bytes) compression has
already been “done”, in Matt Mahoney’s BARF program:
http://mattmahoney.net/dc/#barf
—it might be a better use of your time to download it, play around with
it, examine how it works, and ask yourself, “But is this really useful?
If I were to try and make a better data compressor, what questions would
I really want to ask?”
--
Ian ◎