Signifikand

Signifikand, often called significand in English, refers to the part of a number that contains its significant digits when a number is expressed in scientific or floating-point notation. The signifikand is paired with an exponent that scales it by a base power.

In decimal scientific notation, a real number is written as N = s × 10^e, where s is the significand and e is the exponent. Normalized form typically requires 1 ≤ |s| < 10, meaning the significand uses the leading nonzero digit(s). For example, 6.022 × 10^23 has a significand of 6.022.

In binary floating-point representations, common in computers, numbers are stored as a sign, an exponent, and a significand. The value is often written as ±(1.f) × 2^e for normalized numbers, where 1.f denotes the binary significand consisting of an implicit leading 1 followed by fraction bits. The precision of the representation is determined by the length of the significand. For instance, IEEE 754 single-precision float has a 24-bit significand (23 stored fraction bits plus the implicit 1), yielding about 7 decimal digits of precision. Subnormal (denormal) numbers do not have an implicit leading 1, and their significand is scaled accordingly, allowing representation near zero.

The term signifikand is sometimes used in non-English contexts as an equivalent of significand. It is distinct from the historical term mantissa, which in some contexts referred to the fractional part of a logarithm rather than the significand in a floating-point number.

In summary, the signifikand is the digit-bearing component of a number’s representation, with the exponent determining its scale.