There are two limitations on floating point numbers which are represented in this or any similar way due to the finite space within a computer's memory and circuits:

1. Magnitude limitations -- what is the largest absolute value that can be stored? Fixed by the number of digits in the exponent.
2. Precision limitations -- what is the closest any two numbers can be in terms of size? Fixed by the number of digits in the mantissa.