Numeric parameters (documented as "num" or "numeric") may be of any type supporting numeric values, which includes F (floating point), B (binary), I (signed integer), and even S (string). Note that if the parameter contains or may contain a negative value, you should use either F or I. The size of the variable typically just needs to be large enough to handle the required values. If the parameter absolutely must be floating point or a particular size, it will be documented as such.