Confused with float and double

Question

Hi,
I have done the following program which is supposed to deal with temperature and scales.
import java.math.BigDecimal;
public class Temperature {
     private float temp;
     private char scale;
     public static void main(String args[]) {
          Temperature t = new Temperature(40.5, 'F');
          System.out.println(t.getCTemperature());
     public Temperature() {
          temp = 0;
          scale = 'C';
     public Temperature(float pTemp) {
          temp = pTemp;
          scale = 'C';
     public Temperature(char pScale) {
          temp = 0;
          scale = 'C';
     public Temperature(float pTemp, char pScale) {
          temp = pTemp;
          scale = pScale;
     public float getCTemperature() {
          if (scale == 'C') {
               return temp;
          } else {
               float celciusTemp = 5 * (temp - 32) / 9;
               BigDecimal bd = new BigDecimal(celciusTemp);
               return bd.setScale(1, BigDecimal.ROUND_HALF_UP).floatValue();
}When I try to compile I get the following error :
F:\programs\ch4>javac Temperature.java
Temperature.java:8: cannot find symbol
symbol  : constructor Temperature(double,char)
location: class Temperature
Temperature t = new Temperature(40.5, 'F');
*^*
*1 error*
I don't understand why the number 40.5 is considered like a double and that I can't use Temperature(float, char);
If I'm not mistaken the compiler is asking for a constructor like Temperature(double, char) right ??
thanks.

Answer 1

Can I second Mr TuringPest, and Sir J.Bloch... Do NOT use floats, unless of course you have a really really good reason. Internally, java does all it's floating point arithmetic in doubles anyways, so you're just shooting yourself in the accuracy-foot by forcing it to: promote to double to calculate, "narrow" result back to float... promote, narrow ... promote, narrow ... promote, narrow ... The net result can be astoundingly inaccurate... So just use doubles.