ActionScript as a teaching language

[Lennart Sorensen]

On Sat, Dec 31, 2005 at 05:05:36PM +0200, Peter wrote:
> It looks like Horn clauses. Or like C-notation lambda expressions with 
> ':-', ',', '.' and ';' replacing a lot of parentheses. Or like plain 
> math function descriptions. Which are more or less the same thing. Be 
> f(x) = {x for x>=0, -x for x<0 }. Except the functions can be symbolic. 
> neighbors(A,B,[L]) = { ... }. Writing a simple parser that parses a 
> problem setup as above into valid Prolog is failry easy and could be a 
> part of the default library. More importantly Prolog can explain what it 
> is doing while running. With a simple filter to reduce the verbosity of 
> a trace or explain it should be very helpful.
> 
> I don't think that this is a crazy idea. According to links posted on 
> this list yesterday, using the highest level available language is the 
> best idea. Prolog is certainly high level. For non-symbolic calculus 
> only maybe Matlab or Scilab could be considered.
> 
> If the goal is to teach problem setup and solving, then high level 
> language is good. If the goal is to grind and drill good typing habits, 
> assembly looks great imho, followed by FORRTAN (66) and COBOL probably. 
> If this is an intro course to prepare for a real CS degree then a subset 
> of C++ or something like it should be taught imho, simply because that 
> is the mainstream now.

For something neat that is used in many educational settings, ocaml is
rather nice.  Strongly typed, lots of nice libraries available, ability
to be both compiled and interpreted (interactive even, like python and
basic), and native list types, and partial functions, polymorphic
functions (done right), etc.  Many programing contests for solving
fairly complex tasks efficiently in a short amount of time are won by
people using ocaml or haskell.

Len Sorensen
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml