From: Dan Schmidt <dfan@...>
Date: 2001-02-22T01:40:03+09:00
Subject: [ruby-talk:11245] Re: Summary: RCR #U002 - proper new name for indexes

"James T. Vradelis" <jvradelis@mediaone.net> writes:

| I'd like to see:
| 
|     (k)  Hash#[],  Array#[]
| 
| as in:
| 
|     myHash = {'one' => 1, 'two' => 2, 'three' => 3}
| 
|     myHash['one','two','three']    ->  [1,2,3]
| 
| and
| 
|     myArray = [1,2,3,4,5]
|     myArray[4,2,3]                     ->  [5,3,4]
| 
| I realize that this latter conflicts with:
| 
|     Array#[<beginning index>, <number of elements>]
| 
| but that functionality is duplicated by:
| 
|     Array#[<beginning index>..<beginning index + number of elements - 1>]
| 
| anyway.

You probably already know this, but the Array functionality you
mention already exists:

  [1,2,3,4,5].indexes(4,2,3)             -> [5, 3, 4]

or if you prefer giving it an array:

  [1,2,3,4,5].indexes(*[4,2,3])          -> [5, 3, 4]

Although I agree that the syntax you propose is nicer.

How about the following, which would not be in conflict with the
current overloading of Array#[]?

  myArray = [1,2,3,4,5]
  myArray[[4,2,3]]

That is, arr[anArray] -> aSubArray.

You could even loosen the requirements to arr[anEnumerable], which
would generalize the current semantics for arr[aRange].

I guess this would only work for hashes if you didn't allow Arrays
as keys.

-- 
http://www.dfan.org