Hello World! An Invitation to Computer Science

Lecture Slides

Survey of Programming Languages
Monday March 31, 2008

Recapping the Big Picture

• how to phrase problems in way that computers can solve?
• what are limits of computer v. human brain?
• what are fundamental elements of problem solving of this kind?
• we have identified key concepts:
  • algorithms
  • bits
  • Boolean expressions and circuits (hardware realization)
  • low-level software realization:
      assembly and machine languages
  • key tool at our disposal: abstraction
      lets us zoom/in out on problem
      creates hierarchies of ways of thinking about problem
       so we can tackle large or small problems
       and take context into consideration
• now we get more serious about software realization
    by considering spectrum of programming languages

Assembly language v. machine code

• mnemonics rather than numbers for instruction names
    easier to remember, to read, to understand
    single typo can be flagged as an error
• symbolic addresses
    as with symbolic ops - easier to read, etc.
    no need to change if instruction inserted, deleted, moved
• pseudo-operations for specifying data
• natural language comments

but what are limits of assembly?

Weaknesses of assembly language

• explicit management of all data
• microscopic view
    even "algorithmic primitives" may need to be broken down
• machine specific (not portable!)
• still cryptic: nowhere near natural language
    or standard arithmetic notation

Motivations for high-level languages

• do not have to understand inner workings to write programs
    (is that a good thing?)
• "back in the day": resources were scarce
    major factor in terms of tools/languages available
    and in way people programmed
• positive feedback loop:
    easier-to-use prog langs → more users
      &rarr more demand for computers → more powerful computers
      → more expressive/powerful languages can be developed → ...

3rd generation languages

• addresses above issues (some more than others)
• macroscopic view: allows for abstraction/encapsulation
• require more system software : compilers, interpreters
• libraries, separate compilation, and linking
    increase development speed (see figure 8.1, p. 341)
    programs can be developed using multiple languages

Why are there so many programming languages?

(there are thousands of others aside from those mentioned in text!)

• Tower of Babel effect
• book's analogy of choice among cars
• effect of competitive marketplace
• history of computers and computer science important
• example: polyglot Hello World!

General purpose programming languages

• "Turing complete" (more on that later)
    any program that can be solved via algorithm can
    (in principle) be solved in one a general purpose PL
• important questions remain:
    compactness, efficiency, readability, maintainability, ...
• philosophical debates - can get almost religious!
• "general"; but usually developed for specific
    different PLs better suited for different problem domains
  
  • how fast do you need problem solved?
  • is speed of execution v. speed of development?
  • how much is riding on correctness?
      and how bad is it if an error is found?
  • who else will be reading/using this program?
  • will it have to run on different machines?
      and does it need existing programs or data?
• a crash course in history of high-level PLs

The C programming language

most important language of modern cs (and of last 30+ years)

• intimate relationship with operating system (UNIX)
• portability
• (eventual) standardization
• combination of high-level and low-level features
• C's family: C is ancestor of many popular languages
    very directly: Java, C++, PHP, C#   less directly: Perl, Python, Ruby, Ada, ...

PL considerations

• language features v. library options
• "GOTO deemed harmful" debate
• legacy code and Y2K
• ambiguity for computer v. programmer/reader
• compactness v. mistakes
    C, C++ allow short cuts: hard to read, easy to screw up
    implicit v. explicit distinction among data types

Imperative/procedural PLs

• FORTRAN
• Algol family and Pascal
• BASIC
• C: pointers and exposing low-level details v. abstraction
• C++: C plus much more (object-oriented features)
• Ada - and why it failed
• Java
    why it is so popular?
    why are we not focusing on it in this course?
• C# v. Java

Domain Specific Languages

• SQL (a declarative language)
• markup languages (not programming languages)
    HTML: formatting language for Web browser
    XML: meta language
• Javascript (a scripting language)

Scripting languages

• Perl
• PHP and dynamic Web pages
• Python
• Ruby

Functional programming

• elegant underlying theory
• lists and recursion
• LISP, Scheme, Standard ML

Other programming paradigms

• logic programming
• parallel programming
    SIMD v. MIMD; asymmetric v. symmetric
    distributed v. concurrent v. parallel