From: Dave Thomas <Dave@...>
Date: 2000-11-27T03:12:11+09:00
Subject: [ruby-talk:6581] Re: booleans

Maurice Szmurlo <maurice@europia.fr> writes:

> ruby 3 > a = 10**10000 / 1.5
> ruby 4 > puts a
> Infinity  ## still OK
> ruby 5 > puts a.infinite?
> 1  ## oups! why not "true"?
> ruby 6 > a = 10.0
> ruby 7 > puts a.infinite?
> nil  ## oups! why not "false"?
> 
> (the "finite?" predicate returns true and false)

It's because infinite? is trying to get extra information back to you:

  a = 1.0/0		# => Infinity
  a.infinite?		# => 1
  a = -1.0/0		# => -Infinity
  a.infinite?		# => -1

> conversely if true and false are to exist (which I agree with at
> least for esthetic reasons) why are't they both the only members of
> a "Boolean" class and why isn't nil true (only false would represent
> false)?

I believe a boolean class may be on the way.

Nil is false simply because it's convenient:

   h[key] = 0 unless h[key]

or

   h[hey] ||= 0

or

   while STDIN.gets

and so on.

Personally, I'd argue to remove true and false, as their position is
far more tenuous, but I can live with them once they're here.


Dave