01-31-2003, 04:49 AM
Red_Marvin, your approach may actually work correctly if you try to create the mathmatical formula for blocks of data rather than the entire file. For example if a sequence of bytes in the file were a, b, c, d, e, f, g, ... Then you could create a formula for that block and store a block lookup table at the beginning of the compressed file.
The problem is that not all data can be represented by mathmatical formula, completely random data being the prime example and that it would be computationaly expensive for a compression algorithm to try and derive formulas from data, how would the machine know how big to make the blocks, or the best arrangement.
Most good compression algorithms (ie JPEG and bzip) are hybrids of several compression techniques. A good compression algorithm should also yield good results for most types of data, ie an algorithm that compresses some data very well, but makes random data larger is not ideal.
The problem is that not all data can be represented by mathmatical formula, completely random data being the prime example and that it would be computationaly expensive for a compression algorithm to try and derive formulas from data, how would the machine know how big to make the blocks, or the best arrangement.
Most good compression algorithms (ie JPEG and bzip) are hybrids of several compression techniques. A good compression algorithm should also yield good results for most types of data, ie an algorithm that compresses some data very well, but makes random data larger is not ideal.
esus saves.... Passes to Moses, shoots, he scores!