View source for Webinar 2026 Floating-point Numbers Aren't Mathematical Real Numbers

When one writes computer program code, one might use floating-point values as if they were mathematical real number values. While floating-point numbers are inexact and have a (very) finite precision, they are also not mathematical real numbers. In addition, unlike mathematical real numbers, floating-point numbers are (in general) also neither associative nor distributive. This can result in problems with calculations' numerical stability and correctness using floating-point values. This talk will discuss this and some of the implied issues that one should be aware of when using floating-point numbers in program code.