Browse Books

Reviewer: Teodor Rus

LISP is the first programming language designed on a formal scheme in which a mathematical object was given as the semantics (recursive functions of symbolic expressions) and an appropriate syntax was associated to represent it (s-expressions). LISP has been implemented as a programming system for handling nonnumerical applications. The interpretation technique was usually used for LISP implementation, although the compiling facilities have been added, particularly in the newer versions. It took rather long since its first version proposed by McCarthy in early 1960 until it gained acceptance in the computer science community. Its acceptance is a consequence of the maturation of the need for a formal concept of a programming language required by the necessity of developing a software technology. This book presents a study of the LISP dialects in use today. The declared purpose of this study is, as its title says, to offer to its reader information concerning the performance of the LISP systems, as well as approaches for their evaluation. The mechanism of the benchmarking is used in this respect. However, the author succeeds in covering a larger purpose by collecting into one volume a description of the major LISP systems in use, by providing a set of criteria for their performance evaluation, and by offering a mechanism for understanding the technical tradeoff made during the implementation of a LISP system. The benchmarks used for LISP performance evaluation are LISP packages which have an intrinsic value beyond the measurement purposes discussed in the book. The material is discussed following a methodological approach and is organized on the logical basis in three chapters: (1)A discussion of various computer architectures supporting LISP systems. It provides the theoretical background for performance evaluation and outlines the architectural characteristics that go into the evaluation of a LISP system. (2)The design and implementation of the major LISP systems in use today. (3)A collection of LISP packages used as benchmarks. It discusses the performance of the LISP systems presented in Chapter 2 when executed on these benchmarks. The author emphasizes both the usefulness of a LISP system and its performance. He ends the book with the following conclusion: “The final arbiter of the usefulness of a LISP implementation is the ease that the user and programmer have with that implementation. Performance is an issue, but it is not the only issue.”