Last modified on 22 February 2010, at 10:12

Programming Languages/Introduction

IntroductionEdit

A programming language is an artificial language that can be used to instruct a computer to do a particular task. To be considered a general programming language, it must be computationally complete, or Turing-Complete. It is nevertheless common to regard some languages that are not computationally complete, like database query languages and other domain-specific languages as programming languages as well.

High-Level Versus Low-Level Programming LanguagesEdit

A low-level programming language is one that is very basic and close to the machine's native language. A low-level programming language can be thought of as a building block language for software. Assembly code is the most common low-level language and requires very little translation to assemble it to machine code. (The 1's and 0's that make up binary.)

A high-level programming language is one that is closer to a level of human communication. In this method, the compiler does a lot more of the work for the programmer. The closer the language is to our everyday speech, the easier it is to worry about more complex problems. However, this can be taken too far. If a language is too much like English (or other natural languages), it can be harder to create complex programs. This is because verbose languages take more time to read, and so they can take a lot more time to understand.

Machine code is the language the computer can understand directly. Machine code consists of sequences of binary digits. It is almost never programmed in directly, but anything that is to be run on an ordinary computer must be translated to machine code first. The machine code can be different for each computer architecture.

Assembly language is a more human readable representation of the machine code, where the machine instructions are represented as mnemonics rather than binary digits. Assembly language has a 1:1 relationship with machine code as long as the program is not self-modifying. Before an assembly program can be run by a computer, it must be transformed to machine code. A program that does this translation is known as an assembler. In the early days of computing, assembly language was extensively used, but today it is mainly used for very time critical parts of programs, the core of operating systems, as well as in very small computers, like the chip on a smartcard.

Machine code and assembly language are called first and second generation programming languages respectively. A programming language that has arithmetic expressions, looping constructs, functions, and other constructs that save the programmer from dealing with the machine instructions directly is known as a third-generation programming language.

High-level, domain-specific programming languages were earlier often mentioned as fourth-generation languages, while expert systems were called fifth-generation programming languages. In later years this distinction has blurred, as many very high-level general purpose programming languages like Python, Haskell and Common Lisp have emerged. Expert systems are in very little use today.

Compilation and interpretation of computer programsEdit

Before a program can be executed on a computer, it must be translated to machine code. Alternatively it can be simulated by another program, called an interpreter. A compiler is a program that translates a programming language, called the source programming language into another programming language, called the destination language. Usually the source language is a high level language, while the destination language is machine code. An interpreter may require that the source programming language be compiled into an intermediate form before interpretetion, called byte code. This is a more low level language, for which it is easier to write an interpreter. In the Java programming language this is a separate step, while in other cases it is performed as an integral part of the interpreter. Examples of such programming languages are Perl and Python. CommonLisp is an exception to the above: it's both interpreted and compiled.

Type SystemsEdit

There are two axes to type systems: Dynamic versus Static on the one side and Strong versus Weak.

A Strongly typed language will not allow an operation on an object if this object does not match in type. Examples are CommonLisp, Q-base and Python.

A Weakly typed language will allow such operations. Examples are C and C++.

Dynamic type languages bind type to value. Staticly typed languages bind it to variable.

(See here for background.)

Memory ManagementEdit

  • Manual management
  • Garbage Collection