Floating point normalisation
Ensuring floating point mantissa is in form 1.something (or 0.1something) with adjustment to exponent. Normalisation provides unique representations and maximizes precision. Standard form in IEEE 754.
Formula
Normalised form: ±0.1xxxxx × 2^exponent (mantissa starts 0.1 in binary)
Real World
NASA's Mars Climate Orbiter calculations used IEEE 754 floating point — normalisation ensured maximum precision in trajectory computations where tiny rounding errors over millions of kilometres could mean missing the planet entirely.
Exam Focus
Explain that normalisation maximises precision by ensuring no leading zeros waste mantissa bits; always adjust the exponent when shifting the point.
How well did you know this?