# For extreme accuracy, beyond that of even the double, use strings and do the computations using these single-digit strings. Think about the way you work out general arithmatic by hand then apply it to your script. Theoretically, you could find 21024 to exact precision in considerable time if done properly. The output is a string rather than a variable which allows for this extreme precision.
... and sth like that is what i finally have to do to achieve a correct "str_to_float()" (instead of str_to_num()), i think...
but apart from that it would be cool if str_to_num() would also give back an accurate float if you don't use a var)
