From: Robert Dober <robert.dober@...>
Date: 2009-04-19T18:19:20+09:00
Subject: Re: [ANN] Junctions for Ruby1.9 (Lab419::functional-0.1.2)

On Sat, Apr 18, 2009 at 11:38 PM, Denis Defreyne
<denis.defreyne@stoneship.org> wrote:
> On Apr 17, 2009, at 11:17 PM, Robert Dober wrote:
>
>> E.g.
>>
>>  any(1, 2, 3) > 2     --> true
>>  [ 1, 3, 5 ].all.odd? --> true
>>  none( 1 ).zero?      --> true
>>  all() == nil         --> true
>>  any() == nil         --> false
>
> The examples you gave are equivalent to
>
>        [ 1, 2, 3 ].any? { |e| e > 2 }
>        [ 1, 3, 5 ].all? { |e| e.odd? }
>        ![ 1 ].any? { |e| e.zero? }
>        [].all? { |e| e == nil }
>        [].any? { |e| e == nil }
>
> What advantages does junctions in Ruby provide over #any? and #all? ?
They are an abstraction of the code blocks, if you look at streams,
the second concept implemented in Lab419::functional we have exactly
the same pattern, delay is nothing more than a lambda.
As a matter of fact the main pleasure I have got from releasing code
in Ruby is that Ruby does all the work
and I can get the compliments, well I cannot because you have pointed
out that I have not done anything. Thank you very much ;)

ok seriously now: I believe that this kind of abstraction is useful,
makes code shorter and even more readable.
Is it worth a package? Well maybe not, but what are you going to tell
a perl6 guru if he asks you if Ruby got Junctions? (Proud in Ruby was
indeed a key motivation, maybe that is bad, I dunno )

Look at this example:
  if  a_set.any > b_set.all then
would be
  if a_set.any?{ | x | b_set.all?{ | y | x > y } } then

I kind of prefer to maintain code written in the first style.

Cheers
Robert