The day has come, where 9 is than 9

I was getting red in the face with this thing, trying to figure out
why 9 wasn't equal to 9. Took me a while to figure it out, but I finally found out why. You have a look, it'll make you go nuts if you don't see it at first :p

Code:

---

Private Sub Command1_Click()
'Dim your variables
Dim i
Dim x

'Dim x. Count it with a calculator if you want, but x = 9.
x = 387420489 ^ (1 / 9)

'Finally, do a loop from 1 to 9. Once x = i (Or i = 9), it'll tell you it's equal. But it never does. Aparently, 9 is actually smaller than 9!

For i = 1 To 9
    If x < i Then
        MsgBox "Smaller"
    ElseIf x = i Then
        MsgBox "Equal"
    Else
        MsgBox "Bigger"
    End If
Next i
End Sub

---

Have fun :D

?x-i
-1.77635683940025E-15
Don't use floating points for precise calculations, especially not variants.

The problem is due to round off error when computing the exponential. 1/9 is not exact in either binary or decimal. The exponential function is likely to result in round off error even if the operands are integers.

Quote:

---

Originally posted by Guv
The problem is due to round off error when computing the exponential. 1/9 is not exact in either binary or decimal. The exponential function is likely to result in round off error even if the operands are integers.

---

Exactly. (1/9) cant be calculated exactly in binary...so the computer gets messed up. But it seems the windows calculator actually deals with the fraction so its fine.

Nishantp: A calulator which got exactly 9 for the result probably had some other internal result which rounded to 9 when displayed.

Quote:

---

Originally posted by Guv
Nishantp: A calulator which got exactly 9 for the result probably had some other internal result which rounded to 9 when displayed.

---

I dont know about the Windows Calculator but i know that some calculators are able to do calculations based on the actual fractions so i know its possible (they are probably programmed with methods that we humans use).

X isn't really 9; its like 8.99999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999 9999999999999 or something and VB can't hold that so it rounds up. Either that or X is 9.00000000000000000000000000000000000000000000000000000000001 and the same thing happens.

Fractions don't need special math to calculate them; they are really just divisions that aren't calculated.

Well couldnt you hardcode some functions that do fractional calculations based on human techniques? I wrote something like that to get a huge number of decimals from a division. There SLOW but they work.

Thats pretty much how they work yeh. And they are slow.