Re: [Help-Smalltalk] Metaclass Class model of Smalltalk

[Paolo Bonzini]

On 10/21/2010 02:07 PM, Gwenaël Casaccio wrote:
> Hi,
>
> I want to share an idea about a new model of Metaclass and classes,
> I don't want to refactor gst, but just want to know if it seems good!
> If we look at the class and metaclass model there this infinite loop
> between Metaclass, metaclass of Metaclass, ...
> In fact if we try to build the metamodel this is logic:
>
> metaclass of Metaclass ->  Metaclass ->  ... metaclass of Metaclass ->
> Metaclass ->  Metaclass of ClassOfThatObject ->  ClassOfThatObject ->
> anObject
>
> The problem comes from Metaclass need a class, the class also need a
> class , ... I think you get it and
> this is normal any object in smalltalk has a class.
>
> But what happens if the class of the object  is the object, we get a
> prototype (not new) but this
> is interesting for building a metaclass/class model:
>
> aPrototype ->  Metaclass of ClassOfThatObject ->  ClassOfThatObject ->  anObject

So the Metaclass would be a prototype of itself?  Doesn't it mean that 
"A class class" will see the class-methods of "A", which is wrong?

I think you should think through this more, make an example that works, 
and propose it on address@hidden

Paolo, who's never said "no" so often