On 05/10/11 05:59, Nobuyoshi Nakada wrote:
> Hi,
> 
> (11/10/05 2:14), Alex Young wrote:
>> On 04/10/11 16:52, Nobuyoshi Nakada wrote:
>>> The "flexibility" (or ambiguity) seems the sign that it should not be
>>> in core, to me.
>>
>> I'm not sure I understand.  Where's the ambiguity?
> 
> "Flexibility" can cause ambiguity, sometimes.  Seems it depends on
> application contexts too much.

I don't see that it's any more ambiguous than having Object#===. #==.
#eql? and #equal?.

#nil? has the same relationship to #null? as #== has to #===.  One is a
strict equality comparison, the other is a loose match.  It's
context-dependent in the same way as #=== is: user classes can redefine
it and participate in a global protocol.

>> You could make the same argument about any library.  If something is in
>> ActiveSupport and not in any of the more specifically web-framework
>> Rails libraries, it's also a sign that it's a feature that people find
>> generally useful *outside* Rails.
> 
> Sounds like it's web-framework specific.

I disagree.  Why does it seem that way to you?

> 
>> With respect, I don't find arguments on where a feature has come from to
>> be be particularly relevant.
> 
> Of course, not.  In fact, I'm one of who proposed to introduce
> Symbol#to_proc strongly.  I just said that everything in ActiveSupport
> won't be suitable for the core always.
> 

I'm not suggesting it because it's in ActiveSupport. I'm suggesting it
because it's a generally useful concept.

-- 
Alex