[Home] [Groups] - Message: [Prev in Group] [Next in Group]
8709: [MUD-Dev] Re: PDMud thread summary
[Full Header] [Plain Text]
From: Chris Gray <cg@ami-cg.GraySage.Edmonton.AB.CA>
Newsgroups: nu.kanga.list.mud-dev
Date: Tue, 27 Oct 1998 21:12:31 -0700
Organization: Kanga.Nu
[Alex Oren:]
>} >I think we should look into the object model first.
>} >Do we want single or multiple inheritance? Static or dynamic binding? Single
>} >or multiple dispatch? Vtables or name-based? Etc...
>Static binding is when the destination of a function call is known at compile
>time. Dynamic binding is when the destination is only known at run-time. In
>C++, dynamic binding is accomplished using virtual methods and pointers or
>references to objects.
Static versus dynamic binding of function calls as a programming language
feature I understand just fine. It was those words attached to a question
about an object model that confused me. The object model and the
programming language are strongly intertwined, but they *are* separate.
Anyway, I agree - having a way to attach function pointers (or whatever
you choose to call them) to objects is extremely useful.
>Now, to multiple dispatch. Multiple dispatch is what Scott Meyers (More
>Effective C++) calls "Making functions virtual with respect to more than one
>object". C++ and Java do not support this paradigm directly but there are
>ways to simulate it.
OK, definitely jargon then. So, like, er, the state of two objects
determines which function is called for something?
--
Don't design inefficiency in - it'll happen in the implementation. - me
Chris Gray cg@ami-cg.GraySage.Edmonton.AB.CA