2. 1 The early history : the first programmer 2 The 1950s: The first programming languages

• 
  Early electronic computer
  

• 
  Machine language programming
  

• 
  Assembly language programming
  

• 
  Overview
  

• 
  by John Backus at IBM
  
• 
  scientific and computational programming
  
• 
  new features : array, loop, if,
  
• 
  new versions
  

• 
  Efficiency of execution
  

• 
  efficient code generation
  
• 
  important design goal of FORTRAN
  

• 
  Overview
  

• 
  developed by Algol committee (1958-1960)
  

• 
  a basis for the current imperative languages like
  

• 
  new concepts
  

type declarations for variables

recursion, call-by-value parameters

• 
  Writability
  

• 
  the quality to express a computation clearly, correctly, concisely, and quickly
  
• 
  Example
  

• 
  Overview
  

• 
  by U.S. DoD by a team led by Grace Hopper
  
• 
  business-oriented language
  
• 
  new features
  

- separation of data structures form the execution section

- versatile formatting for output

• 
  Readability
  

• 
  the quality to understand the nature of a computation easily and accurately
  
• 
  Example
  

• 
  Overview
  

• 
  designed at MIT by John McCarthy
  
• 
  function languages for AI applications
  

• 
  Expressiveness
  

• 
  the ease to express complex processes and structures in programming
  
• 
  recursion in LISP
  
• 
  Disadvantage : expressiveness can conflict with simplicity
  

• 
  Overview
  

• 
  PL/I project at IBM
  
• 
  combine all the best features of FORTRAN, COBOL, and Algol60
  
• 
  concurrency and exception handling
  
• 
  Translator
  
• 
  difficult to write, slow, huge and unreliable
  
• 
  difficult to learn and error prone to use,
  
• 
  due to the large number of unpredictable interactions among language features
  

• 
  Overview
  

• 
  improve Algol60 by
  
• 
  creating a more expressive (generality)
  

• 
  The generality principle
  

• 
  Avoid special cases in the availability or use of constructs
  

• 
  Examples of violation
  

• 
  equality operator “=”
  
• 
  can be applied only to scalars, pointers, and sets
  

• 
  no restriction in Ada
  

- Variable length array

- no variable-length arrays in Pascal

- variable-length arrays in C and Ada

• 
  Disadvantages
  

2) make the language less readable and reliable

• 
  Orthogonality
  

• 
  Language constructs should not behave differently in different contexts.
  
• 
  Thus restrictions that are context dependent are nonorthogonalities.
  

• 
  Example of violation
  

• 
  Pascal functions can return only scalar or pointer type (consider assignment in Pascal)
  
• 
  C functions returns values of all data types except array
  
• 
  In Ada, this nonorthogonality is removed
  
• 
  Pascal file types
  
• 
  a special status, for example, files cannot be passed as value, and no assignment to file variable
  

• 
  Overview
  

• 
  designed for simulation
  
• 
  an object-oriented language
  
• 
  class concepts
  

• 
  Complexity control
  

• 
  background
  
• 
  the growing complexity of programming and languages
  
• 
  abstraction mechanism
  
• 
  control the complexity of a programming task
  

• 
  Example
  

• 
  Beginner’s All-purpose Symbolic Instruction Code
  

• 
  Overview of Pascal
  

• 
  Designed by N. Wirth
  
• 
  Simplified version of Algol-68
  

• 
  Overview of C language
  

• 
  Tries simplicity in different ways from Pascal
  
• 
  retaining and restricting the expression orientation
  
• 
  reduce the complexity of the type and runtime system
  
• 
  more access ot the underlying machine
  

• 
  Simplicity
  

• 
  Over simplified examples
  

• 
  BASIC
  
• 
  lacks declarations and blocks
  
• 
  makes it much more difficult to program large applications
  
• 
  Pascal
  
• 
  lacks good string handling, separate compilation, etc.
  

• 
  Consistency of language design
  

• 
  Violation of consistency
  

compute GOTO

GOTO (L1, L2, . . ., Ln), I

assigned GOTO : indirect GOTO

ASSIGN 20 TO N

GOTO N, (L1, L2, . . ., Ln)

ASSIGN 20 TO N

GOTO (20, 30, 40, 50), N

(unpredictable GOTO)

GOTO I, (20, 30, 40, 50)

2) DO 99 I = 1.10 in FORTRAN

3) Pascal

- repeat, while statements

repeat

until (condition);

begin

end;

function f : boolean;

begin

f := true;

• 
  Overview
  

• 
  Abstract data type
  
• 
  Package
  
• 
  Concurrent programming
  
• 
  Task
  
• 
  Exception handling
  

• 
  Question
  

• 
  Overview
  

• 
  Developed by B. Stroustrup
  
• 
  Object-oriented extension of C
  
• 
  Object, class, inheritance, ...
  

• 
  Complexity Control Revisited
  

• 
  Background
  
• 
  the growing complexity of programming (S/W crisis)
  
• 
  Abstraction mechanism
  
• 
  control the complexity of a programming task
  

• 
  Examples
  

C++ : Class

• 
  functional languages
  

• 
  Scheme, ML, Miranda
  

• 
  logic languages
  

• 
  Prolog, Concurrent Prolog, ...
  
• 
  Constraint Logic Programming
  

• 
  Internet programming environment
  

• 
  JAVA
  
• 
  a just simplified version of C++ with powerfule libraries
  
• 
  compile into Byte code of JVM and interpret
  
• 
  executable on Internet environment
  
• 
  portable on Browser like Netscape
  
• 
  Script languages for Internet programming
  
• 
  Perl, Java Script, VRML, …
  

• 
  SML(Standard ML),
  
• 
  Constraint Programming
  
• 
  CLP(Constraint Logic Programming), ...