Hi --

On Tue, 30 Aug 2005, Christian Neukirchen wrote:

> "David A. Black" <dblack / wobblini.net> writes:
>
>> Hi --
>>
>> On Tue, 30 Aug 2005, Christian Neukirchen wrote:
>>
>>> "David A. Black" <dblack / wobblini.net> writes:
>>>
>>>>> Because I felt they are different, as I said above.  I admit fcall is
>>>>> not the best name, but very few people would use it.
>>>>
>>>> I guess I still think of the difference as a kind of mild "danger"
>>>> (i.e., danger to the receiver: its access rules will be violated).  I
>>>> can certainly live without send!  But "function call" just seems so
>>>> out of place.  What about:  send_abs (send absolutely)?  I don't
>>>> know... I'll keeping thinking, if you're still taking suggestions.
>>>
>>> I very much like Austin's idea, to make #send not call private
>>> methods, but keep __send__ doing that.  The latter should not be easily
>>> overwritable either.
>>
>> My problem with that is that the point of __send__ is to keep the name
>> send available for user-defined methods (the same logic that has led
>> to the freeing up of "id" through the creation of object_id).
>>
>> I think adding special significance to the __'s is too much of an
>> afterthought, and also would reintroduce the same problem for people
>> who want to write methods called "send".
>
> No, not really... if you want others to allow reusing "send", you can
> still call __fcall__ (no private).

I'm afraid I'm losing track.  __fcall__ would be the same as send, and
__send__ the same as fcall?  I think I must be getting it wrong.... :-)

I think that in every case throughout the language, if both abc and
__abc__ exist, they should be synonymous.

>>> (And note that a fair number of classes define #send to do something
>>> different already.  When #send and #send! do something entirely
>>> different, ugh.)
>>
>> I'm not sure what you mean.  There is no #send!
>
> Assuming the #send (no private)/#send! (everything) approach was
> accepted, this could get very confusing.

I'm not sure why (one method does one thing; another does another; it
seems very simple), but it's a moot point as Matz doesn't like it.


David

-- 
David A. Black
dblack / wobblini.net