From: Kristof Bastiaensen <kristof@...>
Date: 2004-05-01T22:04:01+09:00
Subject: Re: Lazy evaluation (evil)

On Sat, 01 May 2004 10:38:55 +0000, gabriele renzi wrote:

> il Sat, 01 May 2004 12:29:59 +0200, Kristof Bastiaensen
> <kristof@vleeuwen.org> ha scritto::
> 
> 
>>
>>From a conceptual point of view is callcc very important.
>>It is not just added because it may seem like a good feature,
>>but because it makes Ruby a real functional language (like scheme).
> 
> I think this sentence may be incorrect.
>  I mean, haskell, ocaml and lisp all does not have call/cc yet they're
> functional languages.
> (actually I'm not sure if lisp, ocaml and scheme can be defined as
> functional languages, cause they're not pure ones, if I'm correct)

You are right, I used the term "functional language" in a wrong
way.  Ruby isn't functional, because it doesn't encourage a 
functional programming style.  A meant to say that Ruby treats
functions and continuations as first class objects (just as 
in functional languages).  You could take a scheme program and
rewrite it in Ruby.  It is a pity that the current Ruby-interpreter
doesn't handle tail-recursive call's, I hope that will be in 
Ruby 2.0.  

Another point where Ruby would come short, is in
scoping rules.  Ruby has no equivalent for scheme "let".
Especially with the new local-scoping rules, I have the
impression that it isn't possible to have block local
variables (or am I wrong?).  Wouldn't it be useful to have a
local indentifier, to mean the start of a new scope for 
these variables:

a = "foo"
4.times do |i|
  local a = i * i
end
puts a
=> "foo"

Kristof