Use base 32 for the digits representation

[Laouen]

currently, a number is represented as a vector but each digit goes between 0 and 9. The operations + - * / must be modified to use a higher base that uses the full int number.

Also the << operator must convert the number from the high base to the decimal base.

[sdavtaker]

It may be worth to pass the integer size as template parameter.
You can get it size with sizeof.