
Originally Posted by
Fu Siyuan
Not all files can be compressed to such small. Only high redundancy data have this possibility. Even the outdate WinZip(Deflate) is able to compress to 1/1000 most, however only on extremely high-redundancy file such as the same char repeated milion times.
But I still very doubt that if your observation is right. I don't think 4gb ISO can be compressed into 10mb.
The huffman coding is only a part of such technology. Actually arithmetic coding are used more now. The first document I read is RFC1951. I think it will suit you.