CSE 130 Lecture Notes, October 4, 1999

CSE130 LECTURE NOTES

October 4, 1999

WELCOME

This is CSE 130. Please see today's introductory handout. Points to note:

Very restricted add policy.
Sections on Mondays and Fridays.
Three excellent TAs: Maki Sugimoto, Mark Smith, and Renata Teixeira.
Do come to lectures, but not if you won't pay attention.
Four programming projects, each with a written report.
Teams of exactly three.
Absolute academic honesty is required.
Easy to ask questions with the class bulletin board.
Account slips to be distributed in last 10 minutes of class today.

Class will start at 6:50pm sharp and end at 8:10pm sharp.

OVERVIEW OF THE COURSE

A programming language is a notation for writing software. Good programming languages have many properties:

Readability for humans
Portability across multiple OSs and multiple CPU architectures.
Abstraction, i.e. letting the implementation decide details instead of the programmer.
Performance.
Programmer productivity, especially facilitating code reuse.
More...

A common theme is that the design of the PL should help the programmer avoid mistakes. To err is human!

All the properties above are design objectives. Sometimes the objectives can conflict, for example portability and performance. But often multiple objectives point in the same direction.

Why study programming languages?

To understand the most widely used languages (e.g. C++) better.
To choose the best language for new applications.
Perhaps as a foundation for research on improved languages.

Because PLs are interesting!

HISTORY OF PROGRAMMING LANGUAGES

Most of the ideas of modern PLs appear in four or five classic languages:

Fortran (1956/8): Jump-based control structures (looping, conditionals), subroutines, arrays, formatted I/O.
Cobol: Task-specific types for business applications, e.g. decimal arithmetic and strings.
Algol (1960+): Lexical scoping, composite types (records), automated type-checking, high-level control structures, recursion.
Lisp (1959): Functions yielding functions as return values, same notation for code and data, dynamic typing, recursion in lieu of iteration.
Simula (1967): Information hiding, object-oriented programming.

I said "four or five classic languages" because Cobol has not had much impact on later languages, although it remains important.

Fortran's most important innovation was the very idea of a high-level PL, i.e. that a compiler could generate machine code from a program written in a machine-independent notation. But CSE 130 is not about compilers.

CSE 130 TOPICS

We will study modern languages that are in widespread use:

C: Close to standard sequential hardware.
Pascal: Designed to minimize programmer errors, and for structured flow of control.
ML: Very high-level, first-class functions, no destructive assignment, no explicit pointers, automated type inference.
Java: For safe, portable, distributed computing; object-oriented.
Ada: For embedded, fault-tolerant computing; full type-checking.
awk: For rapid prototyping of text-oriented applications; dynamic typing.

We won't study C++ directly, because it is not as well-designed as the languages above. It is an interesting sociological question why C++ is so widely used today. Possible answers include that C++ is widely available, close to traditional CPU architecture, and familiar to most programmers since it is based on C.

CSE 130 will be organized by concept, not by language. All the languages mentioned above have many more similarities than differences. We will concentrate on the core concepts that all widely-used modern languages are based on.