That's because the map gets translated into the code, this way you can translate the code into the map. So if a map contains many blocks it takes more code to describe the type of block, orientation and position of every block.
