What I meant was: Just because the temperature has a decimal point does not mean you need a float. Multiply every value with 100, so 44.3 degrees becomes 4430. That is enough precision and range for any practical purpose (excluding nuclear power plants and astronomy 
)