From: Ashley Roeckelein <ashley.DONT@...>
Date: 2001-05-22T16:11:25+09:00
Subject: [ruby-talk:15532] Re: Newbie GC question

Mathieu Bouchard wrote:

> On Tue, 22 May 2001, Ashley Roeckelein wrote:
> > OK, thanks for the reply matz.  I will stick with symbols for now...  I
> > like the symbol notation better than the string notation anyway, the :
> > being kind of like the ' in Lisps.
> 
> It's closely like the # in SmallTalk. It is a very constrained version of
> Lisp's quote: it is a quote that only applies to symbols. In constrast,
> Lisp's quote can be wrapped around any kind of expression, transforming it
> into a literal:
> 
> (+ 2 2) returns 4
> '(+ 2 2) returns (+ 2 2)

Yep.  "..kind of like...", but not quite the same :^)  Internally, my 
program uses arrays to represent Prolog clauses, and symbols as both 
variables and constants (I distinquish variables by their initial uppercase 
letter or underscore).  Example: [[:likes,:X,:Y],[:likes,:Y,:X]]  (which 
corresponds to  likes(X,Y) :- likes(Y,X). in Prolog)  This is a little bit 
easier on the eye and pinky than [['likes','X','Y'],['likes','Y','X']].

> The equivalent in ruby would be to add something like a %e{} block, which
> would return instances of subclasses of RubyNode, if those classes were
> visible from Ruby programs, which they are not and won't for the
> foreseeable future.

That would be cool, one day!  Neat stuff could spring from exposing more 
internals.

> > FYI, my "learn Ruby" program is a Prolog-in-Ruby (I know I know - good
> > lord, why?), and it is coming along nicely!  I am very impressed with
> > Ruby's powers so far!  Bravo!
> 
> Please, please, please publish this once you're done or half-done.
> 
> matju
> 

I plan on it!  I'll announce it here (if that's appropriate to do - gotta 
read a faq)

-ash