>>>You're missing my point. You fail to show that your rule is better.
>>
>>Your rule is simply incorrect... There was a thread about the same problem in Visual FoxPro Club in Russian about month ago. You may try to find it.
>
>I'm sorry Nadya, but it's not enough to say that it's is incorrect, you've to prove that.
I've just posted a test result in reply to Elmer. Anyway, here it is:

local i, ab, lnSec
lnSec =seconds()
for i=1 to 10000000
    ab=1
next
?seconds()- m.lnSec     
clear all
lnSec =seconds()
for i=1 to 10000000
    m.ab=1
next
?seconds()- m.lnSec

14.115sec. vs. 16.743sec.