Principles of Programming Languages Review

1
Principles of Programming Languages Review

• April 1, 2001

2
What did we learn?

• Part 1 principles
• Formal languages and automata.
• Scanning and parsing
• Attribute grammars
• Part 2 another point in the space of prog.langs.
• Scheme and Common Lisp.
• Part 3 fundamental aspects of prog. langs.
• Variables, scope and bindings
• Expressions, statements and control
• Data types, garbage collection
• Subroutines
• Building a runnable program
• Data abstraction and object-orientation

3
Formal languages and automata

• Why did we need this material in this course?
• We only scimmed the surface.
• There are more powerful automata (Turing
  machines).
• There is an elegant theory that describes the
  difficulty of problems in terms of whether they
  can be solved
• at all by an algorithm (decidable)
• in a meaningful amount of time and/or space
  (NP-complete).
• CSCI 4112 (Theory of computation) covers the
  theory

4
Scanning and Parsing

• The first stages of analyzing the program to be
  compiled.
• Based on formal languages and automata with a lot
  of pragmatic extensions.
• Generating scanners and parsers is usually done
  automatically from a formal specification (reg.
  expr. or CFG, see lex/yacc)

5
Attribute grammars

• Extend the power of context free grammars
• To capture semantics
• To generate lower-level code
• They are a hack cheating on the formal grammar
  and automata theory.
• But they are useful and necessary.

6
Scheme

• Why introduce Scheme?
• A different model of programming language
• Typeless variables
• Run-time type checking
• higher-order procedures
• closures as first class objects
• continuations

7
Fundamental Tradeoffs

• C/C and Scheme/Common Lisp at two opposite ends
  of the spectrum,
• All other languages somewhere in between.
• Tradeoffs in terms of
• efficiency
• safety
• expressive power.

8
Fundamental aspects of programming languages

• With the previous material as a basis, we moved
  on to examine several fundamental aspects of
  programming languages.
• variables, storage, scope and bindings
• expressions, statements and control
• data types, safe pointers, garbage collection
• subprograms
• constructing runnable programs.
• data abstraction and object-oriented prog.

9
Approach

• The approach was to
• Briefly sample the design choices of various
  programming languages.
• Consider implementation issues at a high level.
• Discuss the fundamental tradeoffs of efficiency,
  safety and expressive power.

10
How to introduce programming

• Various schools of thought
• Start with an industrial strength prog.lang.
  (e.g. C/C)
• Start with a conceptually clean prog.lang. (e.g.
  Scheme)
• Go from low-level (built-in data and control) to
  higher level aspects (like data abstraction) --
  objects last
• Go from higher level (data abstraction) and embed
  low level firmly in it -- objects first

11
How to learn languages beyond the first one

• By identifying similarities and differences to
  the language(s) you already know.
• This is independent of how you were taught your
  first language.

12
What did this course try to accomplish?

• To give you a birds eye view of programming
  languages, so as to speed up the process of
  learning new languages.
• Side effect to focus on what is important and
  what is secondary in a language.
• To recognize that apparent quirks in a language
  may be due to design choices in the fundamental
  tradeoffs.

13
What this course was not.

• A programming course
• A compiler design course
• A software engineering course

14
How else could this course have been taught?

• As a more shallow encyclopedic review of
  programming language features
• As a compiler design course
• Using a fully implemented interpreter as the
  basis for discussing prog. language features (see
  book by Friedman, Wand, Haynes)

15
How this course relates to other courses

• Functional and logic programming focuses on the
  functional programming style, lambda calculus,
  and logic programming. More emphasis on these
  unorthodox programming paradigms.
• Object-oriented programming focuses on
  object-oriented programming, more emphasis on
  software engineering issues.
• Software engineering focuses on the
  process/tools/management of developing large
  programs by teams of programmers.

16
Conclusion

• I hope that you learned something useful in this
  course, and that you enjoyed the process (to the
  extent possible).