CO620

When a number is represented as a continued fraction, then it comes with a natural error bound. Continued fractions can be expressed as digit streams. Arbitrary precision can be achieved by truncating the stream appropriately. Introducing more terms will refine the representation whilst preserving the ability for further refinement. The value of continued fraction arithmetic has been recognized by the functional programming community, because continued fractions can be naturally implemented as lazy streams, but is not as widely known in logic programming. Delay declarations can be used to orchestrate the control needed to compute numeric results lazily to the required degree of precision. Irrational numbers can be represented by infinite continued fractions, which, if they have recurring patterns, can be represented exactly by rational trees. This project demonstrates how continued fraction arithmetic works and how it can be implemented using logic programming features to achieve the desired precision of a result. 1