Re: virtual fn, destructor

Martin Drautzburg <Martin.Drautzburg@web.de> writes:


  An object-oriented language does not need to have classes
  or inheritance, but it needs run-time polymorphism (also
  known as ?late binding?).

  In my opinion, the main advantage is the possibility to extend
  verbs without the need to change existing code. This supports
  the open-closed principle for verbs. (I use ?verb? to mean
  names of functions or operators here.)

  A related feature of C++ is the possibility to overload ?<<?
  for custom classes. In a procedural language, there only would
  be one single procedure called ?operator <<?. To extend ?<<?
  for another type, one would need to /modify/ this procedure
  definition, which is error-prone or sometimes not possible (if
  the definition is in a second-party library).

  But, IIRC, in C++, ?<<? has static dispatch? Now imagine the
  same thing with dynamic dispatch and you got what I believe to
  be the core feature of an object- oriented language.

  A similar opinion is:

      ?This is the heart of why ObjectOrientedProgramming is better.?

http://c2.com/cgi/wiki?ReplaceConditionalWithPolymorphism

                             ~~

  The term ?object-oriented programming? was coined by Alan Kay.
  In 2003, I asked him, what it means and he kindly replied to me:

      ?OOP to me means only messaging, local retention and
      protection and hiding of state-process, and extreme
      late-binding of all things. It can be done in Smalltalk
      and in LISP. There are possibly other systems in which
      this is possible, but I'm not aware of them.?

http://www.purl.org/stefan_ram/pub/doc_kay_oop_en