Forum - Improve Art Studio. Ideas on file compression.

ZX Review #7-8-9-10
(C) Fes'kov Kuzma, Abakan


   For the column "Forum" I propose today advice on improving 
the well-known to all program ART STUDIO, more precisely, its 
version for 128 kilobytes. 

   Remember how many agonizing minutes you spent in trying to
mention two of the same window. On
my opinion, make the move the window a few pixels on a bad 
decision, but in this situation there is one simple solution. 
When this option is selected the arrow changes to a square, 
whose center corresponds to the corner. I propose to replace 
this box with "+", whose center and will be the starting point 
for the window. The cursor will look like when

editing under a magnifying glass. Not
try to find a sprite, corresponding to this square,
all the more complicated (in my
opinion, the authors have wasted a couple of tens of bytes). So 
what also need to change:


   After the program started, we must change the following 
bytes - 57,569.16, 57,572.16; 57575.16, 57578.254, 57581.16;

57584.16, 57587.16. This can be
done using the built-in
ROM monitors. Owners of other
computers will have to find the address yourself.


   I want to share with you
their ideas about file compression. In particular,
method of Hoffman. At this stage,
and indeed for the Spectrum, the idea
archives it seems to me you do not want
because the size wrong, and
programs now are compressing the authors. And Need
self-extracting files (such as those that produce compressors 
KSA and others). 

  To select for the files containing the data of one type, a 
certain tree decompression me seems a little awkward, because 
even for these files difference can fluctuate wildly. My method 
of constructing the tree is built decompression simply, the 
truth, and the compression of a file takes some time (and not 
so great). Before the beginning of compression, we must build a 
tree, on which will be compressed, and this is done very simple 
- we just must count the number of identical bytes, that is, 
how much Once the file meets the code 0, 1,

2 and so on (here we mean not in a row following bytes, and
All occurrences). Next, we
should sort the data descending. This is done
pretty simple. The obvious advantage of this method is the 
possibility of constructing a tree reduce its size. This, in

particularly true for text files, where the size of the tree
sometimes reduced to 100, and then
and 150 bytes, whereas in the proposed size of your way
always fixed. In simple
programs (even very large)
As a rule, you can find 10.5
bytes that do not fall into
They never.

   I've shared just an idea, and maybe it will help someone. 
Lead resulting from I have a program I did not, so

as it is far from perfect,
yes and decompressor is too slow.


           *