What's the universal formula for compression algorithm?

[Q_Me]

Does anybody know? There has to be a universal law of compression of data no matter what the algorithm.

Like if you wanted to compress 01100010 (98) and 00010111 (23) into less or equal to 8 bits, what will that algorithm be when 1 or more numbers contain a different algorithm of doing it?

I can't get it out of my head. Somebody help me out here.

[sql_lall]

i'm not sure exactly what u mean, but two things:

1) These are 8 bits already, yet i guss you mean 'less-than 8 bits'
2) If you consider all 2⁸ possible bytes (8-bit words) then there is no 'universal compression' to change them into the 2⁷ combinations of 7-bit words, such that you can decompress all of these words back to their 8-bit meanings.

The only reason why stuff like winzip work is cos some combinations basically beacuse some longer strings of bits practically never appear.

[Q_Me]

I already knew that, I'm talking about if there is a universal law of compression that will always equal out to what its base(s) were.

What I'm saying is, if you have over 16 strands of 8bit data what is the formula to compress and decompress into a smaller strand or to the original base, even if you don't know the algorithm but you have a piece that equals out correctly to it.

For example, it's kinda like codeword crossword puzzles; you only have 1 or 2 letters to start out with, but you have to fill out the rest of the crossword using only real words but starting with only 1 or 2 letters.

I know how compression works and I know binary code, I just need to know if there is a formula to figure things out.

[jim mcnamara]

Read up on Huffman coding and LZW. Then you'll know why what you are asking doesn't 'compute'. It can't be generalized.

http://www.rasip.fer.hr/research/com.../fund/huffman/

http://dogma.net/markn/articles/lzw/lzw.htm