CSC302 2005S, Class 2: A Brief History of Languages (and computation)

Admin:
* Homework a for Friday: Read Gries, 0 and 1
* Note on readings
* Homework b for Friday: Explore course Web and answer questions
* Sam's frequent absences
* "Quiz": Dating ideas
  + 1. Algorithms
  + 2. Devices to aid in computation
  + 3. The general-purpose computer (as opposed to ...)
    - That is, input, CPU, memory, output
  + 4. Punch cards
  + 5. The first programming language
  + 6. The first implemented programming language
  + 7. Hypertext

Overview
* Some historical discussion of key ideas (see above)
* Sam's biased perspective on the development of programming languages

* Some historical discussion of key ideas (see above)
  + 1. Algorithms
    Def'n: A set of instructions that permit us to compute output (a result) given particular inputs
    Late bce
    Al'Khorizm
  + 2. Devices to aid in computation
    Late bce: E.g. abacus
  + 3. The general-purpose computer (as opposed to ...)
    - That is, input, CPU, memory, output
    von Neumann - 1940's
    Babbage - late 1800's
    3b. First implemented general purspose computer?
      Early 1950's?
      ABC - Atanasoff Berry Computer (Iowa State, early 1940's)
      Harvard-Aiken Mark I - Early 1940's
      Competing British claim
  + 4. Punch cards
    Jacquard loom, mid-1800's, Weaving through punched cards
    Census use in late 1800's
  + 5. The first programming language
    Lady Ada Augusta, Countess of Lovelace, designed language (and programs) for Babbage's computer
  + 6. The first implemented "high level" programming language
    FORTAN - 1954-57 or so
  + 7. Hypertext
    Early 1940's - Vannevar Bush proposed Memex (Atlantic Monthly)
    Bush's article prompted Doug Englebart to invent the mouse and much of other modern stuff
    Talmud is an early hypertext:
      A series of writings and commentatires and criticisms on the hebrew bible

The twelve-minute history of programming languages

* First computers: Programmed by rewiring
* von Neumann suggests strored-program model
* First stored-program programs written as sequence of bits (machine code)
  - Programmers responsible for memorizing bit-codes
  - Programmers resopnsible for laying out data
* Assembly code: Mnemonics for instructions and data, programmatic conversion to bits
* Problems with assembly code
  - Requires lots of attention to details, such as memory locations
  - Programmer responsible for lots of extraneous stuff (such as hardware)
  - When you switch computers, you get a new assembly language
* A team at IBM led by John Backus said "We need a higher-level language that can be easily translated to different assembly languages" (FORTRAN)
  * Yeah, right, those programs will be too slow to be worthwhile
  * A key goal of FORTRAN was the generation of programs "as good as those written by a competent assembly language programmer"
  * FORTRAN was a success!  Generated good code, relatively quickly, significant increase in outcome for programmer effort

Four models of "what is a computer good for?"
* Complex numerical calculations: FORTRAN (Imperative)
* Management of data: COBOL (Imperative)
* Simulating intelligence: LISP (Functional)
* Communicating algorithms to each other: ALGOL (Imperative)

All late 1950's

Object-oriented programming was developed in mid 1960's (Simula-67)
Declartive programming (logic programming) in early 1970's (Prolog)

Every modern language is, in essence, a variation of one of these models with a little bit better software engineering in the design.