| Richard 2006-01-20, 9:55 pm |
| > So, looks like to me, in the extreme cases, zipping a file/record results in
> LARGER chunks than the original.
You say 'file/record' as if they are the same thing, but they are not.
'zip' is a file archiver that includes the ability to compress the data
files that it archives. Because it is an archiver then the zip file
will contain headers that describe the file with all the naming
information, dates, sizes etc and then will have the file data. This
is a file archive. When compression is used the file data blocks take
up less space (usually) than the file itself does.
If you were to use a file compressor that was _not_ a file archiver,
such as Unix compress, or GNU's gzip, then the resulting compressed
file is smaller than the original because it does not include all the
directory and archiving information that zip adds.
for example:
> ls -al xyz
-rw-r--r-- 1 riplin riplin 96 Jan 21 07:31 xyz
> gzip xyz
> ls -al xyz
-rw-r--r-- 1 riplin riplin 66 Jan 21 07:31 xyz.gz
When compressing a record one is not archiving a file, so the file
header information is not needed, just the result of compressing the
data area.
In other words the zip program is not the correct tool for the job of
showing how record compression may or may not work.
|