🚀 go-pugleaf

>>> On Mon, 23 Oct 1995 15:44:42 GMT, shang@corp.mot.com (David
>>> L. Shang) said:

David> C++ and SmallTalk uses a different inheritance mechanism.

Amazing news from America once again...

David> In SmallTalk, the member defined in superclasses actually exists
David> in superclass. SmallTalk uses dynamic inheritance,

Are you sure? This seems to make little sense to me. As far as any
user/implementor of Smalltalk knows, an object is a single entity, not a
collection of entities, as many as there are classes in the prefix
chain.

David> therefore, "this" and "super" are different objects.

No, no, no, no. I read on page 37 of "A little Smalltalk", a slim
introductory level textbook by T. Budd:

  "Both the [pseudo] variables 'self' and 'super' refer to the
  receiver of the current message. When a message is passed to
  'self' the search for a method [ ... ] On the other hand, when a
  messag eis passed to 'super', the search begins with [ ... ]'

In other words, in Smalltalk 'self == super' is true by definition, and
it is only method lookup that differs. The different behaviour is
possible because the Smalltalk compiler understands 'self' and 'super'
to be pseudo-variables, and thus treats them specially.

David> In C++, there shouldn't be a distinction of objects: an object of
David> subclass and an object of superclass. It is a single object
David> viewing in different types: in the superclass and in the
David> subclass.

Exactly like in Smalltalk...


BTW, you are the second person in a day to display a gross
misunderstanding of Smalltalk fundamentals -- and this is not by chance,
or your fault alone. One of the big problems with Smalltalk is the
extremely misleading termonology used therein, e.g. 'message' for
'procedure' and 'sending' for 'calling'. As D. Hopwood reminded me once,
no amount of pithy verbiage will remedy conceptual pitfalls, but I can
try :-), even with languages where the terminology does mislead.