Bug#214925: ent addoption
This certainly seems like a very interesting program to me. I was actually
looking at writing something like this for work.
I am not a Debian developer, and I haven't looked at this program yet.
I've looked at the upstream website and it looks like once I get my hands
on the code I'll be looking for sponsor.
It looks like there's room for some wishlist items including changing
output that says "Optimum compression would ..." as there is no such known
thing, it's likely calculating optimum Huffman compression. Also, this
program does it's calculations at the byte level (one byte per block)
which has benefits and drawbacks. There is an option for bit level which
is probably more useful, but changing block size and even having some
tests for variable sized blocks would be useful. Huffman has been shown to
be optimal for fixed sized block substitution, but BWT and PPM based
algorithms have been shown to produce better compression using other
techniques... Actually this is listed as a bug on the author's website.
Fwiw, there's two Lintian warnings, one for the standards version, one for
a depreciated symlink.
At first glance I'd want to prepare an upload to fix the Lintian warnings,
change the maintainer and fix the upstream bug by changing "Optimum
compression" to something more appropriate.
Drew Daniels
Reply to: