double and float loosing precision - java core knowledge

[Manuel Moons]

Hello,

          double adouble = 50.12333;
          System.out.println(adouble);
          adouble = adouble - 50;
          
          System.out.println(adouble);
          System.out.println("float ----------------------");
          
          float afloat = 50.12333f;
          System.out.println(afloat);
          afloat = afloat - 50f;
          
          System.out.println(afloat); 

The output of this program is:

50.12333
0.12333000000000283
float ----------------------
50.12333
0.12332916


Can anyone explain these values? I would expect no loss of precision with such small values!

[Jeroen Wenting]

value doesn't matter when you're dealing with floating point arithmetic. There's no magical number after which suddenly you start to loose precision, no magic formula that will tell you if and if so how much precision will be lost at any point.

http://docs.sun.com/source/806-3568/ncg_goldberg.html