1989 · Theory & algorithms
Floating-point arithmetic needs carefully designed rules as well as fast hardware.
Visual guide
Use a finite set of numbers.
Apply defined arithmetic rules.
Track the numerical consequences.
Predictable numerical rules make scientific software easier to analyse, test and move between systems.
Standardised arithmetic does not eliminate rounding error or rescue every unstable numerical algorithm.
01 · The idea
William Kahan helped make floating-point computation more dependable. Representable numbers have gaps, so arithmetic must round, and small choices can have large consequences. His work in numerical analysis and the development of IEEE floating-point standards helped establish consistent behaviour for a wide range of machines.
Predictable numerical rules make scientific software easier to analyse, test and move between systems.
02 · The people
William Kahan studied how finite-precision arithmetic can mislead otherwise sensible numerical calculations. He developed algorithms and analysis that expose rounding errors and reduce their consequences. At Berkeley, his work connected mathematics, software, and the design of floating-point hardware. He played a major role in shaping the IEEE floating-point standard, helping make numerical behavior more predictable and reliable across different computing systems.
Read the biography ↗03 · Read further
Original explanations based on the official award record. The subject label is an editorial navigation aid.
Complexity theory connected random choices with some of computation’s deepest limits.
The story
Numerical libraries connect mathematical algorithms with changing hardware.
The story