INTRODUCTION TO IMPERATIVE LANGUAGES

1. DEFINITION

QUESTION: What is an imperative language?

ANSWER: We can define such languages according to the characteristics that they display:

• By default, statements (commands) are executed in a step-wise, sequential, manner.
• As a result order of execution is crucial.
• Destructive assignment - the effect of allocating a value to a variable has the effect of destroying any value that the variable might have held previously.
• Control is the responsibility of the programmer - programmers must explicitly concern themselves with issues such as memory allocation and declaration of variables.

QUESTION: Out of the five principal programming language paradigms the imperative is the most popular, why?

• The imperative paradigm is the most established paradigm.
• It is much more in tune with the computer community's way of thinking.
• Imperative programs tend to run much faster than many other types of program.
• Programmers are prepared to sacrifice some of the advanced features and programming convenience generally associated with higher level languages in exchange for speed of execution.

However, on the downside there is a tendency for programmers to get "bogged down" with control issues.

QUESTION: Compiled or non-compiled?

• Most modern imperative languages are designed explicitly for compilation.
• Older imperative languages (e.g. BASIC and APL), and some very high level imperative-languages (e.g. ICON) are interpreted.

2. EXAMPLE IMPERATIVE LANGUAGES

All modern imperative languages can trace their origins back to three imperative languages: FORTRAN, ALGOL 60 and COBOL. Consequently these influential languages are often described as foundation languages.

FORTRAN The IBM Mathematical FORmula TRANslating system.

• Designed circa 1955 (by, amongst others, Backus).
• First successful attempt to improve on "assembly languages".
• Still widely used for numerical applications.
• Has been revised to take account of ideas on structured programming, however is decreasing in popularity.

ALGOL 60 ALGOrithmic Language 1960.

• Developed in 1950s through a joint European-American committee.
• First block structured language.
• First language whose syntax was defined using BNF.
• Direct ancestor of most modern imperative languages.

COBOL COmmon Business Oriented Language.

• Developed in 1950s through a committee consisting mainly of US computer manufacturers.
• Designed to process large data files and is therefore the most extensively used language for data processing.
• Since its initial introduction the language has remained broadly unchanged.

Relationship between a number of common imperative (and other) programming languages is shown in the "wiring diagran" presented in Figure 1 (click on individual languages for more detail!).

Figure 1: Relationship between a number of common imperative programming languages.

A large number of modern imperative languages follow the style orginated in Algol 60. These lanaguages include Ada and Pascal, and can thus be described as Algol style langauges. There is also a significant group of languages that can be described as C style languages, these include the Object Oriented language C++ and Java. The most widely used imperative language found in industry is currently (the year 2000) C, which tends to dominate the market (with C++ and Java catching up!). This is thus one of the languages that will feature in the following pages. So that we can compare C style langauges (and in particularly C) with other imperativeclanguages we will also feature Ada and Pascal which belong to the Algol style of programming language. Pascal is still a popular teaching language, while Ada was for many years the language of choice for U.S. ministry of defence (arguably the worlds largest procurer of software) contracts.

3. FEATURES OF IMPERATIVE LANGUAGES

• They are usually "typed". Broadly we can identify two categories of imperative data type:
  1. Basic data types.
  2. Compound data types.
• Components comprise:
  1. Data declarations.
  2. Expressions which yield values.
  3. statements which carry out some operation, e.g. assignment statements, conditional statements, program constructs.
• I/O and error handling mechanisms.
• A method of grouping all of the above into a complete program - (program composition).

4. STRUCTURED AND MODULAR PROGRAMMING

4.1. STRUCTURED PROGRAMMING:

The 1960s saw remarkable developments in the field of electronics with the result that more and more powerful hardware components could be produced more and more cheaply. In parallel it was assumed that it would similarly be possible to construct ever larger and more complex software. This assumption proved to be quite wrong - programs failed to be ready on time, greatly exceeded their budget, contained many errors and did not fulfil customer expectations. This phenomenon became known as the software crisis. Among the reasons for this crisis was poor project management, and the fact that many programmers considered their programs to be their property. Thus many individual and curious programming styles developed, and it proved difficult to create error-free programs. In order to remedy this, the concept of structured programming was devised in the late 1960s (by Dijkstra and others). The aim of structured programming is to create programs that are:

• Understandable and hence maintainable), and
• So far as possible, error free.

Structured programming can be said to be a set of rules and recommendations for how "good" programs should be written. It emphasises programming using sequences, conditions and repetition (not jumps and goto statements). These are all constructs which have a predictable logic, i.e. they are entered at the top and exited at the bottom.