CompuServe Messages

Let me add my 2 cents to the thread. Chicago, a.k.a. Windows 4.0, is essentially Windows for Workgroups with a new user interface: a notebook metaphor much like the Mac has been forever. The reason its not on platforms other than intel is because of its DOS/Windows 3.x heritage. There's just too much intel assembler in the DOS underpinnings of Chicago to make it feasible to port to another architecture. Trust me, Motorola tried very,very hard to make that happen for the PowerPC. Bill himself participated in those negotiations. The outcome was "no", and in case you didn't understand, "hell no." Windows NT represents the first pass at an important client/server operating system for the future. Its current incarnation is a CMU Mach-like kernel with layered system services just like most Unix's. In addition, it features true peer-to-peer networking that is almost too simple to appreciate what it took to make it so easy. Cairo, due sometime in 1995, but more likely in early 1996, will spell the end of Windows/DOS as we know it. At that time the NT kernel will become the new host for the Chicago user interface. This combination and its apparent death nell for DOS may seem a radical move for Msft but think about it: 1. NT with its flat memory model, preemptive multitasking, multithreading, multi-platform support and multiprocessor support gets early adopters moving; 2. Chicago, or what will commercially be Windows 4.0, will introduce a new user interface, then; 3. Cairo, or what will be known as Windows NT 4.0, combines the rich feature set of Windows NT with the notebook metaphor of Chicago. By then, memory will be much cheaper, hard drives will be under 50 cents per megabyte and processors will have the peformance capabilities to manage this complex combination of features and ease-of-use. Its the migration path that matters. Any one product by itself will not make or break a company with a solid technical foundation. As far as the HAL (Hardware Abstraction Layer) is concerned, Windows 4.0 only supports intel, so there's no need for abstracting the underlying system from the OS internals. As with DOS, Chicago will own the system and its devices. Take it from a small company who wrote their own HAL, its no small task but well worth the effort. Its as close an approxiamtion to the ANDF (Architecture Nuetral Distribution Format) expounded by the soon-to-be-defunct, Open Software Foundation.