Development of an Interpreter for LRT using the Exact Real Number Paradigm

Abstract

The exact real number representation can grow arbitrarily so it does not truncate or rounds up as opposed to oating point number representation. The advantage of this representation is that not rounding errors are generated and any operation can be achieved with the desired accuracy. The LRT is a proposal that implements exact real number. This paper develops an interpreter for LRT using this paradigm whose operational semantics is based on sixteen rules so the programs based on LRT libraries are less complicated to debug. One of the main problems in implementing LRT has been memory consumption. The main contribution of this work is that LRT has its own administrator for in nite lists as well as its own lazy evaluation. The performance of programs written in LRT interpreter is shown to be superior to the libraries of functions in both execution time and use of memory.