Floating point representation allows computers to store very large or very small numbers using a mantissa and an exponent. This A-Level Computer Science topic covers normalisation, precision and the trade-off between range and accuracy.