CSC302 2007S, Class 03: The Evolution of Some Key Language Features

Admin:
* Today's risk-takers: O'Neil
* Please read The Story of Mel, a Real Programmer for Monday.
* Why is it called Iceweasel?
* Grading time on Thursday eaten by recommendations.
* Homework 2: Permutations, is now available.
* There was some confusion about the readings.  
  They are distributed *two* class days in advance, whenever possible.
* Please learn to use Wikipedia and Google.

Overview:
* Reading Ryder et al.
* Major Characteristics.
* Relating to Past Papers.

/Detour: Programming in C/

Labels: Things that look like variables, but are followed by colons
  L1: fprintf(stderr, "You've reached point L1.\n");
Goto statement:
  goto LABEL;

Hint: If a one-based array index is A[foo], zero-based is A[foo-1]a

/Detour: Tony Hoare/

* Hoare invented a famous sorting algorithm: Quicksort
* Hoare invented a famous control structure: The case statement
* Hoare co-invented CSP, one of the first great models of
  parallel computing

/Reading Ryder/

* What is the thesis?
  * Software engineering has affected the design of modern 
    programming languages.
  * Software engineering and programming languages have mutually
    affected each other.
* What is the structure of the argument that supports the thesis?
  * Interviews with language designers: How did SE support the
    design of your language?
  * Dates of when features seemed to be added and dates of new software engineering design issues came up
    * Need to set up a cause and effect relationship - demonstrate that the SE people were listening to the language people, or vice versa
	  * They were the same people
	  * Citations and references
	* How do we know that something is "a language paper" and that
	  something is "a software engineering paper"?
	  * The focus of the paper might suggest this
	  * The journal or conference in which the paper is published 
	    gives a good indication of how the author thought about 
		the research
	  * Note the disciplinary split that seemed to happen in the early
	    70's with the creation of POPL and some SE conference.

Were they successful?
* The interviews seemed to show less effect
  * Kernighan: Nope, no effect (-)
  * Magnusson: Unsure
  * Wirth: Wasn't called SE, but had an effect (+)
  * Stroustrup (C++): Operating Systems
  * Ada folks: Designed with significant input from SE community
  * Java: Who designed Java?
* Observation: Some bias in the interviews (particularly who)

Are there other biases in the paper?
* Focus on certain features - It may be that these 
* Focus on certain languages - C++, Java, Ada95 (all OO), Ada83,
  a bit on Pascal, and Visual Basic
  * Most focus on large project development
* Other languages in wide use: PHP, Python, (Ruby), Perl
* Other languages that have evolved over time: Fortan, Lisp,
  Java

Detour: Why design and release a language? (Sam will ask again
on Monday)

Big picture: Which of Sammet's predictions are discussed in this
paper?
* Not English as a programming language.
* User-defined languages?  Not really
* Application-specific languages.  paper describes general-purpose
* Fortran will survive about five-ten more years.  paper does not indicate
* Two things that did crop up:
  * Languages better suited for the end user 
    * Some of the interviews
      indicated this (Lindholm claimed that about Java)
	* Abstraction generally supports this
	* Visual Basic was written with this particular audience in mind
  * A more scientific approach to programming
    * The big picture of the paper