Michael,
According to Brad Dryer, one of the SI developers, SI has no deal whatsoever with Microsoft.
ref:
Message 4754021
To Michael Assad:
I agree that home systems are much too difficult to configure. If you consider their evolution, this makes sense, but I agree it must be fixed.
When Apple came up with the Lisa, reintroduced later as the Mac, they had an opportunity to prevent the problem. Since Apple owned the OS, the first applications, and all hardware & peripherals, they built in a sort of simple plug and play with the few devices available. When other vendors tried to offer alternative Mac platforms, such as Franklin, Apple sued to eliminate the competition. This both prevented the growth of semi-compatible systems and peripherals and discouraged OEMs from rushing to build alternatives. This way, Apple could avoid dealing with the incredible combinatorics in a competitive hardware environment.
IBM, on the other hand, had less control of their environment. They licensed the OS from Microsoft in a nonexclusive deal. Since Seattle Computer Products actually wrote the earliest versions of DOS and had visions of becoming a large computer manufacturer, Microsoft wasn't in a position to provide exclusivity (and probably didn't want to). This allowed clone makers, like Compaq, Eagle Computer, Sony, Tandy and others to build semi-compatibles. This, in turn, created a huge market for various peripherals.
Like Apple, IBM had very few options of their own and just built their few optional drivers into the operating system. Unlike Apple, IBM had an incredible amount of 3rd party hardware support & various device configurations. PC/MS-DOS didn't even pretend to fix the problem. If you had a new device, you just had to pop open the config.sys add a couple lines with config parameters, edit the autoexec.bat and you were done. Horrendous. Of course to make it "easier", all the various hardware and application vendors had an auto install that creamed your config.sys and autoexec.bat, replacing it with a new world view.
Then the next problem started to unfold. In order to innovate in applications, companies started including systems functionality in apps or utility software. PC/MS-DOS didn't have functionality to support multitasking, resident programs, fancy keyboard or user input hooks, or graphics APIs. Some of these applications became extremely popular, people started demanding more OS features (like the European only release of multitasking DOS 4.0), and the industry had a problem.
At the same time, Microsoft was trying to build a useful GUI. They failed at Windows 1.0, they finally had a little functionality with Windows 2.0 & 2.1, but Windows still lacked a lot of important OS functionality. Another thing holding back Windows was the fact that all the popular applications wouldn't get along with each other or Windows. The major lesson Microsoft learned out of 2.x was that they couldn't leave behind ISV applications while striving to create a new, more functional system.
About that time, Microsoft bought a company called Paradynamical systems. The people at that firm had a pretty strong OS background and brought some great technical expertise to the Windows team. Until that time, people inside Microsoft thought that the only way they could get the real OS features customers were asking for was to start from a totally new architecture. Systems architects like Gordon Letwin and Dave Cutler focused on operating systems with new base architectures while the Windows group built 3.0. With Win 3.11, Microsoft finally made a usable, network ready, personal OS, not ready for home use or enterprise servers, but usable for typical computer users of the time. NT started to also become usable & more robust than Win 3.1 with the same interface. Still, no real home system.
It could be argued that the Mac was ready to be a home system at that time, but it was really lagging in OS features. Release after release, Apple kept promising to deliver a kernel (you know, multitasking, dynamic memory allocation, virtual memory, etc.), but they never did. Macs were easy to use, but if Windows were light weight in OS features, Macs were ultra-lite, more expensive, and a more difficult development platform.
With Win95, Microsoft tried in earnest to fix the configuration problem and created Plug & Play. This was a huge undertaking, probably greatly underestimated by people outside Microsoft. I actually believe that plug and play made a huge difference in ease of system configuration. Many of my friends, my Parents as well, could finally use a Windows computer.
At the same time, the registry, which had a flawed architecture and was commonly abused, became a database way beyond its original design. Applications that used it for configuration started running into many problems including a lack of system management. It's easy to criticize the registry and its evolution. I do it all the time, but I also understand the technical constraints, goals, and mistakes that led to its current architecture and usage.
Of course, the home user configuration problem was quite different from the enterprise administration issue. Microsoft is just starting to catch up with heavy duty systems (like IBM, Sun, HP, etc.) on enterprise admin, but there's still a bunch of work to do.
The Web changed things in a very profound way. Of course, it formed the beginning of the Neuromancer scenario with everyone connected. It also made distributed applications accessible and useful for the individual user. Most real distributed applications at the time came from the Unix & mainframe environments. In Unix and mainframes, the distributed model was typically thin client, 3 tier systems. They scale well, and they're cost effective because they minimize the need for duplicated, expensive hardware and minimize client cost. They also traded off client caching, configuration, and rich user experience. In fact, PC applications are still widely in use and people expect the millions of features in what detractors call "bloatware". While most people don't need all of the features of any one productivity application, most people appreciate some set of the more obscure functions.
Enter HTML (simplified SGML), scripting, ActiveX, and Sun's view of Java. I say Sun's view of Java because to Sun, Java is all things. It's not just a language, as Microsoft views it, it's a component model, an operating system, or as Miko, Sun's ex-evangelist called it, indescribable. In Microsoft's view, Java and ActiveX are not mutually exclusive. All of these ways of enriching the client had different levels of success with functionality and configuration issues. Many problems still remain. As Sun works to make Java the only solution to the problem, they have to reinvent solutions to problems that Microsoft has dealt with on the client and PCs, much the same way that Microsoft reinvents solutions to enterprise problems in its growth upward.
So now we are where we are, and there's still no fully functional, magically configurable home system with an appliance simple interface. I believe that we're on a path towards that, but I don't blame Microsoft, IBM, Apple, or Sun for where we are. I do believe that Sun's current vision is flawed. I also believe that Microsoft stands an excellent chance of making the best home & enterprise systems. Obviously, I believe this based on my knowledge of Microsoft's perspective as an employee and Sun's perspective as an outsider. I'm also sure that I'm in the minority opinion on this thread, but isn't that what free discussion is all about?
Thanks,
Mike |
