From: "trans.  (T. Onoma)" <transami@...>
Date: 2004-12-15T01:15:44+09:00
Subject: Re: Array#at(selector,...)

On Tuesday 14 December 2004 10:17 am, trans.  (T. Onoma) wrote:
| Is there good reason to have two different methods for this?
|
|   array.at(index)
|   array.values_at(selector,... )

To help answer my own question. A problem arises with output on single element 
results:

  a = [1,2,3,4,5]
  a.at(0)         #=> 1
  a.values_at(0)  #=> [1]

The only alternatives I see for unification would be to have #at take an Array 
or ignore nil:

  a.at([0])         #=> [1]
  a.at(0,nil)       #=> [1]

But barring that (or a better solution), my REAL PROBLEM boils down to how 
best to implement a method related to this but that doesn't have the same 
difficulty. For example (completely contrived):

  class Array
    def part(n=nil)
      if n
        values_at(0..n)
      else
        at(0)
      end
    end
  end

Is it better to have two methods instead --an #at_part and a #values_at_part? 
Personally I prefer the single method but that seems to go against the 
at/values_at grain. So that's why I'm inquiring into it.

| And then
|
|   array.delete_at(index)
|
| Can't this take selector instead?

This still seems reasonable.

T.