Programs often treat uncertain data as exact values, which leads to unreliable decisions. A GPS reading, for example, is an estimate with error, not a fact.
Uncertain<T> implements the approach described in “Uncertain<T>: A First-Order Type for Uncertain Data” by James Bornholt, Todd Mytkowicz, and Kathryn S. McKinley (ASPLOS 2014). Its evidence-based conditionals ask how likely a condition is, such as whether a speed exceeds a limit with 95 percent confidence, before the program acts. The repository has installation instructions and examples.