Hi Xing,

This is due to precision errors with floating point numbers and there is really no way to avoid this in programming.  Ideally, when using a FOR loop with the ZPL, you should only use integers for the start, stop, and increment.  If you need a decimal number inside the loop, then you should calculate that number inside the actual loop and not rely on a fraction coming from the loop counter.  For example, if you want to step through the normalized pupil coordinates with 11 steps, you FOR loop should look like:
 

numSteps = 11  # this should be odd
minus1 = numSteps - 1
FOR i = 0, minus1, 1
   py = (i - minus1) / minus1
   PRINT py, ", ", 
NEXT

This will print -1.0, -0.8, -0.6, -0.4, -0.2, 0.0, 0.2, 0.4, 0.6, 0.8, 1.0.  There will be no ambiguity when to start & stop the loop and you get your decimal values you want calculated inside the FOR loop.