| Stefano B 2008-02-10, 7:58 am |
| Hello Ray,
regarding the technique I agree with those proposed by Ars=E8ne. In fact
the filesystem architecture
does not allow you to create a file that has a fraction of a byte as a
dimension, so you will have
to someway fill up your file up to the closest bit. This is usual in
compression but not usually
a problem as compressing files of at least various KB this affects
file size minimally, even if in
your example this vanishes the bits you would save.
I would also like to underline that unfortunately a technique of this
type cannot be useful for
compression (you haven't specified your goal, but considered the
argument of the group...).
This is because you cannot compress ANY file to a smaller size;
technically because
if the entropy is maximal, compression limit is the original file size
itself. In simpler terms,
you can't compress, for example, any file of 10 bytes or less to a
file of 9 bytes or less as
there are more possible files of the first type and you cannot find a
1 on 1 correspondence
with files of the second.
Hope this helped,
Stefano
|