Hi again, same question.

Hi again, same question.

Post by a__l__k_ » Sat, 27 Dec 2003 16:18:46


Thanks of all of you who helped.

I need any data compressor that compress the orgiginal source file to
tightly packed sequence of integer numbers. The sequence of separated
integer like
x, y , z , a ,b ,c.... and can decompressed to the original file. For
example
324324,65843,123646,23155432,321.... integer are variable length. The
compressor should pack as tightly as possible.

if all of the integer above converted in binary format, we will need
length
information for each integer coz each integer is variable length.

For Exanple Elias Code
0 X
10 XX
110 XXX
1110 XXXX

so with the integer 16, 15, 7

1110 1111
1110 1110
110 111

or 1110111111101110110111

I wish to find a compressor that did the transform/compress as what I
mean above.

Regards,
Thanks
 
 
 

Hi again, same question.

Post by David A. S » Sun, 28 Dec 2003 01:37:13


XXXX@XXXXX.COM (Ray) wrote in



It interesting you don't define the problem yet you are telling
people the solution. Is this home work or what? Why don't you at
least tell the range of numbers. Is it 1 to some N. Are some numbers
zero or negative are all integers equally likely? These one would
need to know to write a bang up coder.
However you could convert to a bijective format tha is packed totally
dense. Meaning on decompression you would get a unique file start with
a number and followed by one or more comma separated numbers ending
in a number with no comma following. In fact here is how
table one for first read
9 =1111
8 =1110
7 =011
6 =101
5 =001
4 =110
3 =010
2 =100
1 =000

Since only 1 to 9 exists

next table you allow 0 to 9 plus next set 1 to 9
if the next character 0 to 9 you use it if ",digit"
you use next set if no character you end file bijectively

Table for rest of file
,9 = 11111
,8 = 11110
,7 = 01111
,6 = 01110
,5 = 10111
,4 = 10110
,3 = 0011
,2 = 1101
,1 = 0101
9 = 1001
8 = 0001
7 = 1110
6 = 0110
5 = 1010
4 = 0010
3 = 1100
2 = 0100
1 = 1000
0 = 0000

The compressed length max would 1 bit for each comma n-1 commas and
4 bits for each digit used. So if you had 100 numbers that in the number
comma format have 1000 character total it would compress to 99*1 4*(901)
3703 or 462.875 bytes. It would be packed dense but if you want better
go with a bijective arithmetic of some form.

Of course you could go with some Universail numbering like you mentioned
in which case you just write the Elais code for the number the only tricky
thing is the last number you have to use partical Elais for last number
I would use fibanocci (bad spelling ) but that just me.


It really should take you no more than a day or to for
eithe approach. Is this homework,


David A. Scott
--
My Crypto code
http://www.yqcomputer.com/
http://www.yqcomputer.com/
http://www.yqcomputer.com/
My Compression code http://www.yqcomputer.com/
**TO EMAIL ME drop the roman "five" **
Disclaimer:I am in no way responsible for any of the statements
made in the above text. For all I know I might be drugged.
As a famous person once said "any cryptograhic
system is only as strong as its weakest link"

 
 
 

Hi again, same question.

Post by a__l__k_ » Sun, 28 Dec 2003 09:37:43

David A. Scott" < XXXX@XXXXX.COM > wrote in message news:< XXXX@XXXXX.COM >...


It is great ! your idea give me new flash. Thanks Scott.
you helped me alot.. this is not a homework. It is a research.

erm.. data compression research. Not the kind like Jules G. (you know)

Anyway. Thanks again.


Regards,
Raymond