From: Dan Sugalski <dan@...>
Date: 2002-09-12T15:45:31+09:00
Subject: Re: Multimethods (was Re: Larry Wall's comments on Ruby)

At 8:52 AM +0900 9/12/02, Yukihiro Matsumoto wrote:
>Hi,
>
>In message "Multimethods (was Re: Larry Wall's comments on Ruby)"
>     on 02/09/11, Dan Sugalski <dan@sidhe.org> writes:
>
>|Multimethod dispatch is reasonably simple. It's when you have
>|multiple methods with the same name, which are only differentiated by
>|their method signature. At runtime, the types of the arguments are
>|checked and you dispatch to the matching method. If no method
>|matches, or more than one method matches equally well, an exception
>|is thrown.
>
>In my definition, this is not multi-method, because a method still
>belongs to certain class.  I call this "overloading method".  I'm not
>against overloading method.  This has power of multi-method, and
>simplicity of ordinary method.  Just like mix-in compared to MI.

Fair enough. I use the terms Damian used as he was the person who 
explained this stuff to me.

There can also be some argument as to whether something that doesn't 
live in a particular class can be a method in the first place, though 
from my perspective if you can call it as a method it is a method, 
even if where it's stored internally might be someplace else. But, 
then, I'm generally not the right guy to ask about theoretical rigor. 
:)
-- 
                                         Dan

--------------------------------------"it's like this"-------------------
Dan Sugalski                          even samurai
dan@sidhe.org                         have teddy bears and even
                                       teddy bears get drunk