Since the term "Mantissa" can refer to the "Fractional part" of the number for logs, and it can also refer to the "Integer part" and "Fractional part" of the number (combined, without the "Exponent part") for numbers in scientific notation and floating point... it is an ambiguous term and should be avoided.
"Significand" is also not appropriate since it also refers to the "Integer part" and "Fractional part" of the number (combined, without the "Exponent part"), for numbers in scientific notation and floating point.
I prefer to use the terms "Integer digits" and "Fractional digits" (or "Integer part" and "Fractional part").
As far as the method to capture the "Integer digits" and "Fractional digits" for a negative number. Given a negative number like n = -2.3:
(Perhaps this is not important to you because your numbers (data) may all be positive numbers).
Method 1:
While it may be correct from a purely technical or academic standpoint to split this up as:
"Integer digits" = (-)3
"Fractional digits" = (+).7
It may not make sense for you depending on how you will use it.
If you will be treating these parts of the number, also as numbers (rather than "Strings"), and you will at some time combine these two number parts back into the original number, this method has the advantage that you can simply add the two parts of the number together to get the original number back: (-)3 + (+).7 = (-)2.3.
Method 2:
You could get the same effect by storing the sign of the number with each part of the number:
"Integer digits" = (-)2
"Fractional digits" = (-).3
This will also allow you to simply add the two parts of the number together to get the original number back: (-)2 + (-).3 = (-)2.3.
But, perhaps your purpose of breaking the number up is to facilitate displaying the number in a particular way. Neither of these methods would be very useful for this purpose, particularly if you were storing the number parts as strings. Storing the number parts using the first method would take some odd Mathematical gymnastics to get a printable version of the number back.
My recommendation is Method 3:
Split the number up like this:
- Given a number "n" like n = -2.3 or n = 2.3
- Store the "Sign" of the number:
s = Sgn(n)
Or as boolean: s = (n >= 0)
- Remove the "Sign" of the number
n = Abs(n)
- Save "Integer digits" portion:
i = Fix(n)
- Save "Decimal digits" portion:
d = n - i
Or as "String": d = Mid(CStr(n - i), 3)
Or as "Integer": d = ((n - i) * 10000)