Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

can you compress zip files?

0 views
Skip to first unread message

Kevin P Chugh

unread,
Nov 5, 1999, 3:00:00 AM11/5/99
to

sorry if this is a faq but is it ever possible to compress a zip file?
i tried a few popular compressors but they all result in files
bigger than the original. just wondering if zip is the smallest
you can get.

thanks,
kevin

DI Michael Schindler

unread,
Nov 5, 1999, 3:00:00 AM11/5/99
to
yes - you unzip them and then compress the files with a better
compressor. don't expect too much - 5% is all you probably can get

Kevin P Chugh <ch...@cs.buffalo.edu> wrote in article
<7vtiht$l41$1...@prometheus.acsu.buffalo.edu>...

J Shane Culpepper

unread,
Nov 5, 1999, 3:00:00 AM11/5/99
to Kevin P Chugh
Compressed data is always more difficult to compress than uncompressed
data. Its a classical entropy problem. Try converting the zip file to
stored mode (infozip -0) or uncompress the zip file and recompress the
data with something better like rar, imp, bix, tar+bzip, etc. There are
alot of programs out there that compress much better than zip. See
www.act.by.net for alot of comparisons and links. As always, check out
the newsgroup faq at www.faqs.org for more links and info.

Later,

-Shane

Kevin P Chugh wrote:

> sorry if this is a faq but is it ever possible to compress a zip file?
> i tried a few popular compressors but they all result in files
> bigger than the original. just wondering if zip is the smallest
> you can get.
>

> thanks,
> kevin


Mathew Hendry

unread,
Nov 5, 1999, 3:00:00 AM11/5/99
to
On 5 Nov 1999 03:22:37 GMT, Kevin P Chugh <ch...@cs.buffalo.edu>
wrote:

>sorry if this is a faq but is it ever possible to compress a zip file?
>i tried a few popular compressors but they all result in files
>bigger than the original. just wondering if zip is the smallest
>you can get.

Applying normal compression algorithms to a zip file usually won't
help much. But you could use a specialised algorithm something like
the following

1) Identify the version of zip used to encode the file (probably
difficult).
2) Separate "side" info (filenames, security permissions, other file
flags) from the compressed file data.
3) Extract the files
4) Place side info and zip version id in another file
5) Compress the whole lot using a better compressor (bzip2, szip, ace,
...)

If you identified the zip version correctly, and collected all side
information, it should be possible to completely reconstruct the
original zip.

-- Mat.


Sascha Zapf

unread,
Nov 5, 1999, 3:00:00 AM11/5/99
to

Kevin P Chugh schrieb:

> sorry if this is a faq but is it ever possible to compress a zip file?
> i tried a few popular compressors but they all result in files
> bigger than the original. just wondering if zip is the smallest
> you can get.
>

> thanks,
> kevin

Zip is not smallest you can get, but is is very fast for good compression
result. The reason why the result are bigger is that all normal compressor
are looking for the same properties...Run lenghts, patterns thats repeat,
are some of them. If you have another compression algorythmn, can get it
smaller.

Sascha

Message has been deleted

Phil Norman

unread,
Nov 15, 1999, 3:00:00 AM11/15/99
to
On Mon, 15 Nov 1999 12:04:31 -0500, Xiaohui Xue <x...@julian.uwo.ca> wrote:
>0-order arithmetic coder can almost always compress zip files. Although
>the "improvment" is actually too slight, it shows the difference between
>statistics-based compression and dictionary-based compression.

Not really - zip deflate includes statistics-based compression
aswell (huffman) What you're probably compressing is the zip
file headers, and directory information.

Cheers,
Phil


0 new messages