From: Brian Candler <B.Candler@...>
Date: 2004-10-08T20:39:18+09:00
Subject: Re: Range behavior (Re: [RCR] New [] Semantics)

On Fri, Oct 08, 2004 at 05:31:41PM +0900, Yukihiro Matsumoto wrote:
>   planA - leave member? and include? as they are; they are very easy
>           to distinguish, and to remember.  why bother?

In this case, also change the documentation of Enumeration to state that
whilst Enumeration#include? and Enumeration#member? do the same thing, that
they are intended for different purposes (and in particular, that
Range#include? is different from Range#member?)

That for me is the biggest problem - confusion caused by the fact that I
expected them to be the same, because Enumerable says they are aliases.

>   planB - making member? as alias to include?; they behave same.  no
>           confusion, no problem.  who wants membership for ranges?

Hmm, actually that's a pretty good idea IMO. Does anyone care to keep
   (0..3).member?(2.5)  # => false

? I doubt it. There's almost always a direct (non-interative) way to test
that anyway with a Range; nobody wants to iterate for hours to discover
   (0..100000000).member?(2.5) # => false

Remember the problem which triggered all this off, was testing ranges with
infinite bounds...

OK, I think that gets my vote.

Cheers,

Brian.