From: Alex Gutteridge <alexg@...>
Date: 2007-10-10T15:07:03+09:00
Subject: Re: will #in? be in ruby2? (was Re: Finding shared elements between to arrays.)

On 10 Oct 2007, at 14:06, Peņa, Botp wrote:

> From: Devi Web Development [mailto:devi.webmaster@gmail.com]
> # > compare:
> # >    a3 = a1.select { |e| e.in? a2}
> # >
> # I really don't like that proposal. It's not a property of e
> # that it is in a2, its a property of a2 that it contains e.
>
> pls enlighten my poor brain :(
>
> i just want to ask if e is in a2. same way like i want to ask eg if  
> e==2  ...  in fact, i find #in? so generic...
>
> or maybe you want a keyword only like msbasic's
>
>     e in a2
>
> ?
>
> kind regards -botp

If I understand you correctly the :in? method would have to be  
defined in Object (so that all Objects could be tested). So something  
like:

irb(main):001:0> class Object
irb(main):002:1>   def in?(enum)
irb(main):003:2>     enum.include?(self)
irb(main):004:2>   end
irb(main):005:1> end
=> nil
irb(main):006:0> 1.in? [1,2,3]
=> true
irb(main):007:0> 4.in? [1,2,3]
=> false

Daniel's point (I think) is that this pollutes the top level Object  
with a property which for many (most?) Objects is never relevant. If  
you never put it an Object into a collection then the idea of being  
'in' something makes no sense.

On the other hand it is always an intrinsic part of an Enumerable  
implementing class (by definition a collection of Objects) that you  
may often need to know whether it 'includes' a specified object.

I'm not sure why the includes construct rocks your brain though. Both  
read very logically to me (as a native English speaker admittedly):

a3 = a1.select { |e| a2.include?(e) }

reads as: 'select' from a1 those e where a2 'includes' e.

a3 = a1.select { |e| e.in? a2}

reads as: 'select' from a1 those e where e is 'in' a2.

That said I always write this as:

a3 = a1 & a2

Alex Gutteridge

Bioinformatics Center
Kyoto University