Inexact Conversion Result

[MrBcx]

USING$ to the rescue:

Code Select Expand


DIM a![10]

FOR INTEGER i = 0 TO 9
    a![i] = i + 0.1
    PRINT USING$("#.#", a![i])
NEXT

I know Robert knows this, so the following is for anyone else who might be interested in this thread.
The precision issue that we see is related to how floating-point numbers are represented in binary.
For example, the value 1.1 cannot be exactly represented in IEEE 754 binary floating-point.   Our
compilers convert that to the closest representable double, which is 1.1000000238418579.   
No amount of formatting can recover the "intended" 1.1 because that information is already lost,
so we artificially compensate with functions like USING$()

[Robert]

Here's another "Watch out for !" example of inexactitude.

Code Select Expand


DIM a![10]

FOR INTEGER i = 0 TO 9
  a![i] = i + 0.1
PRINT a![i]
PRINT STR$(a![i])
?
NEXT



Result:

Code Select Expand


 0.1
 0.100000001490116

 1.1
 1.10000002384186

 2.1
 2.09999990463257

 3.1
 3.09999990463257

 4.1
 4.09999990463257

 5.1
 5.09999990463257

 6.1
 6.09999990463257

 7.1
 7.09999990463257

 8.1
 8.10000038146973

 9.1
 9.10000038146973