From: "Wolfgang Nádasi-Donner" <ed.odanow@...>
Date: 2007-12-12T23:04:25+09:00
Subject: Re: redefineable not operator

murphy schrieb:
> this is the coolest thing since the ordered hashes ^_^

Well, I want to do some written up thinking and dreaming in
direction of Ruby 2 (or 3?), because I hope that all this
discissions here don't stop December, 24th, 2007 ;-)

When I'm putting together some features of Ruby 1.9, more
operators can be redefined, Unicode characters can be used
freely for method definition (example "real ��" instead of
"lambda"), and when I remember very foggy about features
of Algol68 to redefine operators and priorities (please don't
ask me about details - it's too long ago I used the language),
and when I look at the presentation of Jim Weirich
(RubyConf 2007)...

...I end up with the question "why not thinking about free
definition and redefinition of infix operators and their
priorities for future versions of Ruby?". This includes
wired operators like "&&" and "||", because methods which
fulfil the semantics are easy to write in Ruby...

module Kernel
   def my_and
     (t=self) ? yield : t
   end
   def my_or
     (t=self) ? t : yield
   end
end

require 'test/unit'

class Test_my_logical < Test::Unit::TestCase
   def test_my_and
     assert_equal( 1&&2      , 1.my_and {2}     )
     assert_equal( false&&2  , false.my_and {2} )
     assert_equal( nil&&2    , nil.my_and {2}   )
   end
   def test_my_or
     assert_equal( 1||2      , 1.my_or {2}      )
     assert_equal( false||2  , false.my_or {2}  )
     assert_equal( nil||2    , nil.my_or {2}    )
   end
end

This together with the open classes and modules in Ruby
can lead to a very interesting scenario. We can define
a special Ruby variant in Ruby (a DSL if you like this TLA),
which can be used simply by an tailored (e.g. via .irbrc)
irb from a user.

So - please don't hit me or, even worse, put my sender address
to a spam list. I post this with the intention to produce
some dreams and ideas for versions of Ruby, that are beyond 1.9.x.
I have no idea for an implementation - I didn't think about it,
but for me this here looks like a logical continiation of the
road.

Wolfgang N��dasi-Donner