macOS weird float problem

You're not doing anything wrong. Potentially a float should be able to represent an infinite amount of numbers. After all there are an infinite amount of possible numbers between 1 and 2 for example (1.1 , 1.12, 1.234, etc.). But computers have to use the discrete values 0 and 1 to make a representation of the infinite set of real numbers. So when you try to store 0.1 into a float it will try to store the best approximation. Which in your case will probably be 0.1000000014901161. This is quite a classic computer science problem.

It's also the reason why you should never try to compare a float directly with an other float. Because they might not be what you expect them to be. Always use a certain margin of error.

Try reading this paper about it. What Every Computer Scientist Should Know About Floating-Point Arithmetic. It's quite extensive and maybe even quite complex but it contains some very useful information when you're working with floating-point numbers. There are some other good documents about floating point numbers as well but try some search engine to find them. Knowing about them makes you a better programmer.

Try using -stringValue and -setStringValue and you won't have the problem with the float inaccuracies

caminoix said:

unfortunately, i don't have a mac at home and can't check it till tuesday but it seems to me that it should be quite impossible to use stringValue with float variables - so that i could perform arithmetic operations on them later on, or am i wrong?

Click to expand...

You can easily convert between the two. To get the float from a string use the -floatValue method. and you can convert back using [NSString stringWithFormat"%f", float] (you could also specify the number of decimal places in the string using the @"%.xf" notation). The point is that you keep the values as they were entered by the user until such time as your program needs to do something with them (like comparisons or arithmetic).