This book is about an approach to programming in which simplicity, clarity, and elegance are the key goals. More specifically, it is an introduction to the functional style of programming, using the language Haskell.

The functional style is quite different to that promoted by most current languages, such as Java, C++, C, and Visual Basic. In particular, most current languages are closely linked to the underlying hardware, in the sense that programming is based upon the idea of changing stored values. In contrast, Haskell promotes a more abstract style of programming, based upon the idea of applying functions to arguments. As we shall see, moving to this higher-level leads to considerably simpler programs, and supports a number of powerful new ways to structure and reason about programs.

The book is primarily aimed at students studying computing science at university level, but is also appropriate for a broader spectrum of readers who would like to learn about programming in Haskell. No previous programming experience is required or assumed, and all the concepts are explained from first principles, with the aid of carefully chosen examples.

In this chapter we show how Haskell can be used to write interactive programs. We start by explaining what interactive programs are, show how such programs can naturally be viewed as functions, define a number of basic interactive programs and higher-order functions for combining interactive programs, and conclude by developing a calculator and the game of life.

Interaction

A batch program is one that does not interact with the user while it is running. In the early days of computing, most programs were batch programs, run in isolation from their users in order to maximise the amount of time that the computer was performing useful work. For example, a route-planning program may take start and finish points as its input, silently perform a large number of calculations, and then produce a recommended route as its output.

Up to this point in the book we have considered how Haskell can be used to write batch programs. In Haskell such programs, and more generally all programs, are modelled as pure functions that take all their input as explicit arguments, and produce all their output as explicit results. For example, a route planner may be modelled as a function of type (Point, Point) → Route that takes a pair of points and produces a route between them.

In contrast, an interactive program is one that may take additional input from the user, and produce additional output for the user, while the program is running.

The subject of this book is constraint logic programming, and we will present it using the open source programming system ECLiPSe, available at http://www.eclipse-clp.org. This approach to programming combines two programming paradigms: logic programming and constraint programming. So to explain it we first discuss the origins of these two programming paradigms.

Logic programming

Logic programming has roots in the influential approach to automated theorem proving based on the resolution method due to Alan Robinson. In his fundamental paper, Robinson [1965], he introduced the resolution principle, the notion of unification and a unification algorithm. Using his resolution method one can prove theorems formulated as formulas of first-order logic, so to get a ‘Yes’ or ‘No’ answer to a question. What is missing is the possibility to compute answers to a question.

The appropriate step to overcome this limitation was suggested by Robert Kowalski. In Kowalski [1974] he proposed a modified version of the resolution that deals with a a subset of first-order logic but allows one to generate a substitution that satisfies the original formula. This substitution can then be interpreted as a result of a computation. This approach became known as logic programming. A number of other proposals aiming to achieve the same goal, viz. to compute with the first-order logic, were proposed around the same time, but logic programming turned out to be the simplest one and most versatile.

In parallel, Alain Colmerauer with his colleagues worked on a programming language for natural language processing based on automated theorem proving.