>I didn't know VFP's behaviour like that.
>
>I think so fortran is right... :)


No. The result of fortran is right if you code fortran. The result of vfp is right if you code in vfp.

It is simply wrong to trust assumptions if you deal with formulas and comps.

I do not know but on primary school there was something what we called chain - calculation. It was just to learn basic operaitons

The teacher asks 2+3
Answer 6
T -2?
A 4
T *2
A 8

So nobody would assume that 2+3-2*2 is 8

But this was not the formula. It was
lnResult=2+3
?lnResult
lnResult=lnResult-2
?lnResult
lnResult=lnResult-2
?lnResult
lnResult=lnResult*2
?lnResult

This is the same if you use a four operator caculator or a scientific one.

An other example is the use of MOD(). It looks like Modulo operation but it is not. (comes up twice a year in this forum)