cr88192 wrote:
<denis.r.santos@gmail.comwrote in message
news:1155821911.426609.125490@
>
>for a file of that size, have you looked into bzip2 (libbz2)?
>bzip2 is rather effective for files that fit within a single compression
>block (900kB).
>>
>likely, ppm is not terribly likely to outperform bz2 by any significant
>amount in this case afaik.
>
>
If you go to then you
will see that BZIP2 is well below PPMZ2 in compression performance!!! I
only started this post because I really think PPMZ2 is better (but as I
said I am not an expert in compression, I only look at the practical
results) since I have tested it in a few files and because if it exists
an exe it probably exists, also, a dll associate!! Dont you think ?
--
note that I said 900kB, and the file in the benchmark is 3MB
--
given the way bzip2 is implemented, is is particularly good for files less
than this mark, but not so much so for files over this mark
this is an issue wrt the algo used in bzip2 (the bwt transform), which works
by sorting out the contents of a number of fixed-size blocks.
now, if your whole file fits within a single block, the algo does well, but
if not, some level of overhead is duplicated for every block (because each
is coded independently), and the ratios are no longer nearly as impressive.
or such
--
Thanks for answering cr88192!!!
The best compressors are slow and use lots of memory. PPMd is probably
a good compromise and the source code is free. Here are some more
benchmarks (size, speed, memory) on text.
http://cs.fit.edu/~mmahoney/compression/text.html
-- Matt Mahoney