Thanks for your answers anyway.
Best regards
Karl
-----Original Message-----
Final remark:

That C++ does it wrong in this case does not mean that
the
outcome of a floating point operation is, in general,
exact. It is not. In no compiler. Ever.

Anyway, there are simpler ways of replacing the decimal .
by an underscore.

As about the ignoring rounding errors (or making them
fit,
as you call it): suppose we were dealing with money, why
would rounding the results to 2 decimals be so bad? I
don't know about you, but I can spare 5.7e-12 cents any
time.

-----Original Message-----
You're not serious are you? You mean, as I know the
difference will be 0.1, I should round the wrong result
from the calculation so that it fits 0.1 !!!???

Btw, I'm using this value as a seed for a filename, and
I'm happier if my file is called:
64_1.txt and not
64_0.999999999999946.txt ;-)

I guess you have to be VERY careful with numbers in VB...
Karl
oh, and C++ does it right! (see my answer to your 1st
post)

-----Original Message-----
How to write 64.1-64 and get EXACTLY 0.1 ? (that must
be
possible in VB or not?)

Why? The deviation we are talking about here is of the
same order as 1 cm compared to the distance between the
earth and the sun. If you do not want to see the mess
caused by the calculations, round the result to the
desired number of decimals.
.

.

.