Transcript Milestones in Computing 1940s today

Chapter 2
Evolution of the
Major Programming
Languages
Amid the strife of war...


1936-1945
Konrad Zuse (Tsoo-zuh)
–
–
Built a series of complex computers from
electromechanical relays
Developed a language called Plankalkul for
expressing computations

Unpublished until 1972
ShortCode
A Step Toward Readability/Writability


1949
John Mauchly
–
Developed a language named ShortCode for the BINAC


It consisted of coded versions of mathematical expressions
Example statement: 00 X0 03 20 06 Y0
–
Meaning X0 = SQRT(ABS(Y0))
– X0 and Y0 were variable names; 03 the assignment opearator, 20
sqrt, and 06 absolute value


Machine implementation was pure interpretation, which was
termed automatic programming.
More readable and writable than keying in the machine code
equivalent, but at a cost of running 50 times slower.
1950s
Compiling systems
Relocatable code
Assembly languages
Fortran 0, I, II
AI – FLPL; IPL-I, LISP
Business – Flowmatic & COBOL
ALGOL
Keep in Mind the Initial Context

At the beginning of the 50s
–
–
–
–
–
Primary use of computers were for numerical
calculations
Computer memories were small
Hardware did not directly support floating point
operations or indexing
Hardware was unreliable
Hardware was more costly than programmers
A First Compiling System


1950-1953
Grace Hopper and her team at UNIVAC
–
–
Developed a “compiling system” named A-0, A-1,
and A-2
Programs were written in a type of code called
pseudocode which was expanded into machine
code subprograms
Progress in Generality


1950
David J. Wheeler
–


Developed a method of using blocks of
relocatable addresses.
1951
Maurice Wilkes
–
Extended this idea to design an assembly
program that could combine chosen subroutines
and allocate storage.
Speedcode System
Support for Floating Point Operations


Early 1950s
John Backus
–
–
–
–
–
Developed a Speedcode system for the IBM 701 which
extended machine code to include floating-point operations.
Instruction set included the four arithmetic operations for
floating points, sqrt, sine, arc tangent, exponent and
logarithm; conditional and unconditional branching; I/O
conversions.
Novel facility of auto incrementing address registers.
Pure interpretation (slow -“add” took 4.2 milliseconds)
Faster coding - Matrix multiplication could be coded in 12
instructions
First Real Compiler?


1952
Alick E. Glennie
–
Developed Autocode compiler for the Manchester Mark I computer

Lowlevel and machine oriented
OR


1953
Laning and Zierler
–
Developed an algebraic translation system





Used function calls and included arrays and branching instructions
Implemented on the MIT Whirlwind computer
Generated a subroutine call to each formula, expression
Easy to read
Used within MIT
Fortran
A Significant Step Up – First Important High Level Language


1954
IBM 704
–

Provided hardware for indexing and floating point instructions
John Backus and group at IBM
–
Published a description of Fortran 0
 Major goal: Provide efficiency of hand-coded programs
–
–
–
Note: The first version of the compiler had little syntax error checking based on claims of eliminating
coding errors and thus debugging
Implementation began in 1955
Released in 1957 as Fortran I.








Types and storage for all variables fixed before run time – efficiency vs. flexibility
Included I/O formatting
Variables of up to six characters
User-defined subprograms
If and Do statements
Implicit data types – I-N for integers
18 worker-years of effort
Huge success
Artificial Intelligence
An Influence on Programming Languages

Mid 50s
–
Interest in AI emerged



Natural language processing
Modeling human information storage and retrieval and
other brain processes
Mechanizing certain intelligent processes such as
theorem proving
–
These interests held a common need for processing
symbolic data in lists that could be easily manipulated
 linked lists vs. arrays.
AI and Fortran


Mid 50s
IBM
–
FLPL (Fortran List Processing Language)

Extension to the Fortran compiler
–
Used to construct a theorem prover for plane geometry
First AI Programming Language
IPL - Information Processing Language


1956
Allen Newell, J.C. Shaw, Herbert Simon
–
Published a description of one of the first AI
languages

IPL-I
–
Information Processing Language
The Business Domain


1957
FLOW-MATIC
–
–
Business oriented language for the UNIVAC
“mathematical programs should be written in
mathematical notation, data processing programs
should be written in English statements” – Grace
Hopper 1953
The Business Domain


1959
COBOL 60
–
Common Business Oriented Language

Design Goals
–
–
–

Characteristics
–
–
–
–
–
–
–
Use English as much as possible
Easy to use, even at the expense of being less powerful..
Not be overly restricted by the problems of its implementation
DEFINE verb for macros
Records
30 character names with hyphens
Data division and procedure division
Mandated by the DOD
Example on pages 65-67
Standardized versions in

1968; 1974 (subprograms with parameters), 1985, 2002...
Fortran progresses


1958
Fortran II compiler
–
–
Bug fixes
Independent compilation of subprograms

Made lengthier programs possible
ALGOL
Algorithmic Language



1958
GAMM (Society for Applied Mathematics and Mechanics) and ACM (Assoc. for Computing Machinery)
ALGOL 58 – A Universal Standard Language
–
–
–
Developed jointly by a committee of European and American computer scientists in a meeting in 1958 at ETH Zurich.
IAL (International Algorithmic Language vs. ALGOrithmic Language
The following people attended the meeting in Paris (from January 1 to 16):

–
Alan Perlis gave a vivid description of the meeting: "The meetings were exhausting, interminable, and exhilarating. One
became aggravated when one's good ideas were discarded along with the bad ones of others. Nevertheless, diligence
persisted during the entire period. The chemistry of the 13 was excellent."



–
Friedrich L. Bauer, Peter Naur, Heinz Rutishauser, Klaus Samelson, Bernard Vauquois, Adriaan van Wijngaarden, and Michael Woodger
(from Europe), John W. Backus, Julien Green, Charles Katz, John McCarthy, Alan J. Perlis, and Joseph Henry Wegstein (from the USA).
Design Goals
–
The syntax should be as close as possible to standard mathematical notation, and programs written in ti should be readable with
little further explanation
–
It should be possible to use the language for the description of algorithms in publications.
–
Programs in the new language must be mechanically translatable into machine language.
Characteristics
–
Machine independent
–
Flexible & powerful
–
Simple and elegant
–
Generalized many of Fortran’s features

Identifiers of any length; any number of array dimensions, lower bound of arrays set by programmer; nested selection
statements.
–
Formalized the data type concept
–
Compound statements
–
:= assignment operator
Both John Backus and Peter Naur served on the committee which created ALGOL 60 as did Wally Feurzeig, who later created Logo.
ALGOL 60 inspired many languages that followed it.

Tony Hoare remarked: "Here is a language so far ahead of its time that it was not only an improvement on its predecessors but also on
nearly all its successors."[7]
LISP


1959
John McCarthy and Marvin Minsky
–
–
Produced a system for list processing
LISP





A functional language
Dominated for a decade
Originally interpreted
Example diagrams and program on pages 54-55.
Decendents are:
–
Scheme
– COMMON LISP

Related Language - ML
1960s
IPL-V
ALGOL-60
APL
SNOBOL
Fortran IV
BASIC
PL/I
SIMULA
ALGOL-68
Progress in AI


1960
Newell and Tonge
–
IPL-V


Demonstrated that list processing was feasible and
useful.
Actually an assembly language implemented in an
interpreter with list processing instructions for the
Johnniac machine
ALGOL 60


1960
ALGOL 60
–
–
–
Formally described using Backus-Naur Form
Example on pages 61-62
New additions




–
–
Note that formatted I/O was omitted due to goal of machine
independence
Parent

–
Block structure concept
Pass by value and pass by name
Recursive procedures
Stack-dynamic arrays
Fortran
Descendents

PL/I, SIMULA 67, C, Pascal, Ada, C++ and Java
APL and SNOBOL


1960
APL
–
–
Kenneth Iverson at IBM
Designed for describing computer architecture


–

Implemented in mid 60s
SNOBOL
–
–
–

Odd character set required for operators
Expressivity vs. readability
D. J. Farber, R. E. Griswold, and I.P. Polonsky at Bell Labs
Designed for text processing.
Collection of powerful operations for string pattern matching
Common features
–
–
Not based on a previous language nor a basis for any languages
Dynamic typing and hence storage allocation
Fortran IV


1962
Fortran IV
–
–
–
–

One of the most widely used PLs
Explicit type declarations for variables
Logical If construct
Capability of passing subprograms as parameters
1966
–
Fortran 66 – its Standardized version
BASIC
Let’s make it easy…


1963
John Kemeny and Thomas Kurtz
–
Designed BASIC

–
Goals
1.
2.
3.
4.
5.
–
Beginner’s All-purpose Symbolic Instruction Code
Must be easy for non-science students to learn and use
Must be pleasant and friendly
Must provide fast turnaround for homework
Must allow free and private access
Must consider user time more important than computer time!’
Characteristics





Small, nonteractive
Used through terminals
Single data type – fp – numbers
Resurgence with Visual BASIC in 90s.
Example on page 69.
PL/I
A Single Universal Language


1964
IBM
–
–
Developed PL/I
Goal

Capable of both floating point and decimal arithmetic to support both scientific and business apps, as well as,
support for list processing and systems programming!
–
–
Contributions




ALGOL 60’s recursion and block structure
Fortran IV’s separate compilation with communication via global data
COBOL 50’s data structures, I/O and report generating facilities
A collection of new constructs
–
–
–
–
–
–
Concurrently executing subprograms
Exception handling for 23 different types of exceptions
Allowed the disabling of recursion for more efficient linkage
Pointers as data types
References to cross sections of Arrays
Failings

–
Replace Fortran, LISP, COBOL and assembly languages.
“I absolutely fail to see how we can keep our growing programs firmly within our intellectual grip when by its sheer
baroqueness the programming language – our basic tool, mind you! – already escapes our intellectual control.” –
Edsger Dijkstra 1972
Example on pages 74-75
Simula 67


1967
Kristen Nygaard and Ole-Johan Dahl
–
First developed SIMULA I in the early 60s


–
Designed for system simulation, implemented in mid-60s
Generalized into Simula 67
Features


Extension of ALGOL 60 taking block structure and control
statements
Support for coroutines via the class construct thus beginning
the concept of data abstraction
–
–
–
Encapsulation of data and processes that manipulate the data
Class definition as a template
Constructors
ALGOL 68
Dramatically Different


1968
ALGOL 68
–
–
Design Criteria – Orthogonality
Never achieved widespread use, but contributed several
important ideas.




User defined data types
Flex arrays – Implicit heap-dynamic arrays
Orthogonality – a few primitive concepts and unrestricted use
of a few combining mechanisms
Descendents – ALGOL-W
–
Value-result method of passing parameters as an alternative to
pass-by-name
– Case statement for multiple selection
1970s
Pascal
C
Prolog
Scheme
Fortran 77
Ada
Pascal



1971
Niklaus Wirth
Developed Pascal based on ALGOL 60
–
Primarily used as a teaching language

–
–
Simple but expressive
Lacked essential features for many apps which
led to non-standard dialects such as Turbo Pascal
Example on pages 80-81
C



1972
Dennis Ritchie
Developed the C language
–
Heritage was


CPL - Cambridge early 60s
BCPL – Martin Richards 1967
–

–
B – Ken Thompson 1970 First HLL under Unix
 C
and ALGOL 68
Example on pages 82-83
Logic Languages


1975
Phillippe Roussel
–
–
Described Prolog
Example code on page 86
Scheme
A Functional Programming Language


1975
MIT
–
Scheme




Small size
Exclusive use of static scoping
Functions are first-class entities – can be values of
expressions and elements of lists; assigned to variables;
passed as parameters and returned as values of
function applications.
Simple syntax and semantics
Fortran 77
Continues to Dominate


1978
Fortran 77
–
–
–
–
Character string handling
Logical loop control statements
If else
“Fortran is the “lingua franca” of the computing
world..” Alan Perlis
1980s
Smalltalk
ADA
ML
Common LISP
Miranda
C++
Smalltalk


1980
Alan Kay who predicted computer “desktop”
windowing environment
–
Developed the first language that fully supported
OOP as a part of the Xerox Palo Alto Research
Center (PARC) group



Charged with task of designing a lanuage to support
Kay’s vision.
Objects and message passing
Example on pages 93-94
MetaLanguage
Functional Language Interest Continues


1980s
Robin Milner
–
ML (MetaLanguage)


Functional but supports imperative
Syntax similar to imperative
ADA


1983
DOD
–
Most extensive and expensive language design effort

Climate
–
–
–

More than 450 different languages were in use and none were standardized
Reuse and tools not feasible
High-Order Language Working Group was formed

Id the requirements for a new DoD HLL

Evaluate existing languages to determine candidate languages

Recommend adoption or implementation of a minimal set of languages

April 1975 – produced Strawman requirements docs

August 1975 – Woodenman

Jan 1976 – Tinman

Jan 1977 – Ironman

June 1978 – Steelman

July 1979 – Cii Honeywell/Bull language design was selected.

Spring of 1979 – name was adopted

1983 – “final” official ANSI version
Features
–
–
–
–
Packages
Exception handling
Generics
Concurrency support
LISP


1984
COMMON LISP
–
–
–
–
Designed to combine features of a number of
differenct dialects of LISP that were developed
during the 70s and 80s.
Large and complex
Allows both dynamic and static scoping
Basis is pure lisp
Miranda


1984
David Turner
–
Miranda



Based on ML, SASL, and KRC
Functional, no variables, no assignment statement
Haskell is based on Miranda
–
But has the unique feature of lazy evaluation - No
expression is evaluated until its value is required
C++

Bjarne Stroustrup at Bell Labs
–
–
Made the first step from C to C++ with C with Classes language int 1983
Goals



–
provide a language with classes and inheritance
No performance penalty – reason array index range checking was not considered
It could be used for every application ofr which C was used – so left unsafe features of C
Progression

1980
–
–
–

1981
–
–
1984
–
1985



Addition of function parameter tyep checking and conversion
Classes like those of SIMULA 67 and Smalltalk
Derived classes, public/private access, constructors/destructors, friends.
Inline functions, default parameters, overloading
Named C++ - virtual methods, dynamic binding of method calls to method definitions, reference types
First available implementation named Cfront which translates C++ programs into C programs
Continued to evolve
–
–
–
multiple inheritance, abstract classes
Templates which provide parameterized types and exception handling
2002 - .NET
1990s
Fortran 90
Fortran 95
ADA 95
Java
Fortran 90


1990
Fortran 90
–
–
–
–
–
–
–
–
–
–
Dynamic arrays
Records
Pointers
Multiple selection statement
Modules
Recursion
Obsolescent-features list
Dropped fixed format of code requirement
Fortran vs. FORTRAN
Convention – keywords & identifiers in uppercase
Fortran 95


1995
Fortran 95
–
–
Forall added for parallelizing..
See code example on page 50-51.
ADA 95

Features
–
Adding new components to those inherited from a base
class.
– Dynamic binding of subprogram classes to subprogram
definitions (polymorphism)
– Protected objects

Success hindered by C++
Scripting Languages

Sh (shell) – a small collection of commands
interpreted as calls to system subprograms to
perform utility functions with added varaibels, control
flow statements, functions andetc
–



Ksh – David Korn ’95
Awk – Al Aho, Brian Kernighan, Peter Weinberger
(’88) began as a report generation language
Tcl – John Ousterhout ’94
Perl – Larry Wall – designed as a UNIX tool for
processing text files. Common Gateway Interface
language as well.
Scripting Languages

Perl – Larry Wall
–
–
–
–
Combination of sh and awk
Statically typed variables - $ scalar; @ arrays; % hash
names
Arrays can be dynamic and sparse
Some dangers


If a string is used in a numeric context and teh string cannot be
converted to a number, zero is used without warning
Array indexing cannot be checked since there is no set
subscript range. References to non-existent elements return
undef, which is interpreted as 0 in numeric context.
2000s
Fortran 2003
C#
…
Fortran 2003


2003
Fortran 2003
–
–
–
–
Parameterized derived types
OOP support
Procedure pointers
C interoperability