Talk:Object-oriented operating system
Isn't it WAY to steep to consider Win9x systems front-ends for DOS? If you had said so for Win3.x, I would have agreed, but this sounds like a little too much to me...
- Indeed. I fixed it. The role of DOS in Win 9x is fairly obscure, but can be rougly summarised as follows:
- There are two OSes, real mode and protected mode. The computer starts in one and then switches to the other. The real mode OS could loosely be called MS-DOS.
- Windows APIs often call DOS APIs which in turn call Windows implementations. This is for backwards compatibility. It allows interrupt hooks.
- From 95 to 98 to Me, features have been progressively ported from Windows 3.x era 16-bit code to 32-bit code. Backwards compatibility has been maintained.
-- Tim Starling 12:00 19 Jun 2003 (UTC)
Surely, the concept of an object-oriented OS has nothing to do with the concept of GUI widgets? I have always believed that an OS is object oriented in API design, rather than in GUI implementation.
As far as I can tell, most OSes have OO-like features in their APIs, for instance, in UNIX, everything is a file/stream; in Windows, only handles are passed around; in Darwin, the driver API is implemented in C++; in microkernels, messages are passed between processes (like invoking methods of objects); and naturally, widget-oriented GUIs are object-oriented. (Curiously, X11 isn't -- the only type of graphical object is a window, and apparently, windows don't nest)