McMaster University
CAS 706

Lectures Extra material Languages and compilers Assignements Presentations

Lectures

Tuesday 12:30-13:20 in ITB/222, and Friday 10:30-12:20 in ITB/AB105.

Evaluation

40% assignments, 35% in-class presentation, 25% final exam.

Assignments

Parametrized quicksort. Languages: D, Eiffel, Ocaml, Prolog.
Lambda calculus interpreter. Languages: Lua, Scala, Haskell, mercury.
Typed Lambda calculus interpreter. Languages: Pascal, Javascript, F#, curry.
Fun with scripting languages. Perl, python.
Bonus: implement the AO Challenge in MetaOCaml, where each ``aspect'' should be modularized, in the same style as the modularization for the Gaussian Elimination family. See also this paper.

Textbooks and web materials

Required: Types and Programming Languages Benjamin C. Pierce
A good idea: Programming Languages: Principles and Paradigms Allen Tucker and Robert Noonan.
There is some very useful material at the following pages:

When in the mood to look at weird languages, I definitely recommend

this rather aptly named language
99 bottles of beer in 598 different languages.
Weird and Esoteric languages.

In the same vein, some good jokes based on programming languages.

I recently found a nice comparison of the extension benefits of OO vs functional. Clearly lays out the tradeoffs involved. I could also talk about how FP is (mostly) based on initial algebras and OO on co-algebras, but that might not enlighten many amongst you!

For specific languages, I recommend:

Smalltalk
GNU Prolog
Prolog tutorial
Use gcc for C code. Turn on all warnings.
Haskell. I recommend using Hugs for learning Haskell.
Ocaml. Note that ocaml 3.04 is installed on the CAS servers (in /usr/local/bin); this is fine for getting the assignments done. The latest version of ocaml is 3.07+2, but it is not necessary to use this.
Java. Java 1.4 is already installed on all CAS machines.
You can use either free pascal or use one of the undergraduates machine with Pascal installed on it.
mercury. This is now installed on birkohff (a Sun) and penguin (linux). To run this properly you should
- add /usr/local/mercury-0.11.0/bin to your PATH,
  This can be done (if you use a csh-based shell) via putting
```
 set path = ($path /usr/local/mercury-0.11.0/bin) 
```
  in your .cshrc (in your home directory). Something similar can be put in your configuration if you use a sh-based shell.
- add /usr/local/mercury-0.11.0/man to your MANPATH,
- and add /usr/local/mercury-0.11.0/info to your INFOPATH.
- You may also want to add the following lines to the `.emacs' file in your home directory (if you are an emacs user):
```
        (setq load-path (cons (expand-file-name
                "/usr/local/mercury-0.11.0/lib/mercury/elisp") load-path))
        (autoload 'mdb "gud" "Invoke the Mercury debugger" t)
		
```

All the sites above contain executables and lots of documentation.

Presentations

Topics to choose from (* items need to be covered):

* TAPL Ch 9,11 (typed lambda calculus + extensions)
* TAPL Ch 12,13 (normalization, references)
* TAPL Ch 15,16 (subtyping)
* TAPL Ch 19 (featherweight Java)
* TAPL Ch 22 (type reconstruction - needed for assignment 3)
TAPL Ch 23 (Universal types)
TAPL Ch 24 (Existential types, modules)
* OO: objects, classes, inheritance, subtyping, prototypes
* OO: overloading, overriding and method dispatch
Domain Theory (CPOs, least and greatest fixed points, etc)
Meta-programming
Compiling statically typed functional languages
Compiling dynamically typed languages
Continuations and semantics of goto
Partial evaluation
Operational semantics of Prolog
suggest a topic!

Additional resources

Googling for "introduction to lambda calculus" and "lambda calculus interpreter" yield a lot of useful resources

Dec 2007

McMaster UniversityCAS 706