For-Next loop with non integer values

[charvista]

You understood it correctly Krzysztof

I *think* that Thorium is trying to tell you that computer's floatings is not like we have learned them at school.
For example, what we see as 1.36 can in fact be 1.3599999999998 to the computer. That's why it is good practice to avoid floats whenever possible. It is not a PureBasic error, it is the general mathematical problem with fractions, and it is the same problem in all computer languages, but some languages convert them internally, which make the program running much slower and sometimes erroneously.

There exists a trick to circumvent this and that keeps the accuracy. For money currencies, we often use maximum 2 digits after the decimal point. Transform all values to integer by multiplying by 100 (1.36 * 100 = 136) and perform all calculations with integers in the format "136", and when all calculations are done, show the result divided by 100.

[Thorium]

Yes. As long as you do calculate your values in every iteration in a while loop you get no problems. As soon as you start to increment the value and check it for a expected end result you get problems because you cant increment and compare floats that way. If you absolutly must do it, use fixed point as charvista showed.
For everything else just count with integer and calculate floats from scratch at every iteration, so no incrementation takes place.

[krzysztof]

Thank You very much for all explanations, but this is unclear:

...As long as you do calculate your values in every iteration in a while loop you get no problems. As soon as you start to increment the value and check it for a expected end result you get problems because you cant increment and compare floats that way

You wrote about calculating, incrementing and counting of floats. Could You please, give an example of simple (please!) codes for explanation?

[BorisTheOld]

You wrote about calculating, incrementing and counting of floats. Could You please, give an example of simple (please!) codes for explanation?

Here's the problem with floats.

Code: Select all

x.d = 0.0
While x <> 1.0
  If x > 2.0
    Break
  EndIf
  x = x + 0.1
  Debug x
Wend

[krzysztof]

Very nice and clear example. It explains the problem. Thank You.