>Hi George,
>
>------
>First, it clearly states in the VFP help that precision is limited to 15 digits.
>
>Second, it also states that the cause of this is a limitation with Pentium class processor and follows IEEE standards.
>
>Third, this limitation is not limited to VFP. See MSKB#78113.
>------
>
>Nobody of these affirmations has a tie with this problem.
>
Fabio,

OK, if you say so. The problem is that you're using sixteen digits in the zz variable. Shorten it by removing the last digit to the right of the decimal point and it evaluates correctly. This is consistent with both the VFP docs and the KB article.

If you want to say it's a bug, fine. I bothered to RTFM, and accordingly to that, it isn't and I choose to believe that.