The type of huffman decode table is [[(usize, u8, u8); 16]; 256], where the usize is an index to the next entry within that table. Because the length of the table is exactly 256, wouldnt be storing the index as u8 make the table size significantly smaller ?