The Microsoft Master Plan

By Nils Dahl

Date: May 20, 2002


When Microsoft began the browser war, I wondered just why it was so important to control this application market. I thought it was just some Internet-related notion. I was not thinking big enough.


Then came Active Desktop, a very curious add-on that used the browser to generate the desktop image. It seemed curious to add a layer of processing on top of the existing operating system capabilities.


It all came together when the features of Microsoft's .NET concept began appearing. Truly a neat and elegant idea.


First, imagine that your goal is a worldwide network that extends sideways into local networks in every household, linking together computers, appliance boxes, remote data storage and remote processing, plus all that the Internet has to offer. The easiest protocol to use for tying everything together is TCP/IP with the html/xml packet format.


Imagine schoolrooms where each student has a tablet computer with a wireless link - all hosted by a server somewhere in the building - or perhaps a server/workstation for each teacher. In a similar vein, workplaces can use an assortment of workstations that tie into local servers - which use fiber optic or copper to link into a main server complex.


At home, the main computer links via wireless to gaming systems, boxes that manage broadcast/cable/dish media, roaming wireless tablets that act as controllers, and perhaps even communications boxes that replace wired/wireless telephones or all kinds.


Active Desktop generates a gui display that is easier to broadcast to wireless tablets. The wireless tablet must be a complete system but need only have modest processing power. While VIA's EDEN is an eminently suitable core for such a portable system, it is actually overkill.


Now each home's wireless network MUST have security, if only to avoid intermingling of packets flowing in other homes. The answer is a unique id for each user - via Passport. While this is not foolproof - unless Microsoft could legislate embedding of id chips in all users - it is a mandatory identification scheme. Its not just for online shopper.


Now a student in school can use a wireless tablet, leaving work on a remote server that is easily reachable from the school's servers or from the home server - or anywhere else that is wired in, including shopping malls and stores. Instant messaging lets everyone stay in touch with everyone else - provided that all users behave and that nobody steals an identity. Passport inevitably must be implemented by sophisticated encryption and be stored in flash ram inside all units.


None of this is new. Apple has introduced equivalent technologies starting with its first Macintosh. The Apple Desktop Bus is an early version of usb. The Geoport predates the Winmodem. Apple has wireless systems on the market - the iBook and iMac with base station as well as the more powerful portable and desktop systems with add-ons. Most recently, Apple has announced servers that support large local networks. As usual, Apple defines the technologies and gets them to market first.


Microsoft's idea of using the HTML/XML browser to drive the display image - along with the usual low level capabilities of Windows XP embedded and full Windows XP - is a rather neat approach to offering all the solutions that almost everyone might need or want. Low power portable units are easy to build, once designers realize that flash memory cards are superior to even the smallest hard drives for certain storage needs.


Yes, there are other efficient methods of moving desktop and application interface images around - but Postscript and the other less well-known solutions are all proprietary schemes that require licensing. Ghostscript just isn't sufficiently developed. Certain processing intensive chores such as font antialiasing are moving into the video chips - Matrox's new technology offers this and might be useful enough to move Matrox into a favorable position when lower cost versions of its new designs are released or someone else copies the idea.


For such an ambitious program to succeed, Microsoft needs hardware partners who are willing to work closely with Microsoft. More importantly, controlling costs of supporting a wide variety of networked products means minimizing the cost of coding and recoding product support. That, of course, leads to a conclusion that the firms offering the best x86 code support in their semiconductor products are in the best position to be successful partners. Intel's notion of branching off into new worlds of code and software optimization in new processors that are 'huge' and 'very expensive to make' just don't fit nicely into a concept that runs just fine on x86 legacy code and could even execute nicely on 0.13 micron pentium processors running slow enough to need only passive cooling.


There are obvious potential problems with security, even if a Passport scheme locks all devices in a home into the correct local network. Anything that is designed will, inevitably, be abused or cracked by a few people. On the other hand, imagine libraries where users sign out tablet pc units to use the catalog, access the Internet, use IM and email, or even read eBooks. You can likely imagine many other uses. One of the conveniences I always wanted was access to a database of all the books I own - so that I could avoid buying duplicates AND fill in gaps in certain authors while shopping at a bookstore. The same would apply for music, movies, and many other categories of personal possessions. Buying a very expensive PDA that runs terribly limited software and uses a tiny display screen is NOT a practical solution. Buying an expensive PDA that cannot be upgraded is out of the question.


If Microsoft can establish partnerships throughout the computer industry, its grand plan could extend much farther - into retail stores of all kinds. Instead of Telxon wireless computer terminals, retail people could use color graphic wireless tablets and get far better stocking and ordering information. Imagine going into a video rental store and punching up a trailer for any movie you are considering renting? Or you might listen to music in a store before buying the album. The possible uses are limited only by one's imagination.


It is equally obvious that Linux is a bare bones foundation that would require extensive add-on work and support to achieve the range of ready-to-use forms that Microsoft and its partners can easily produce already. Satisfying mass market is all about delivering easy to use solutions that just work. Even the best of the packagers of Linux are way too small and far too limited in vision to ever hope to compete against Microsoft. Then there is the curious problem of weak or missing hardware support for most add-ons.


While building the pieces of such a system and getting them to work is fairly easy, selling the idea can be very difficult. Big businesses tend to prefer hiring total solution firms that can give them what is needed at a reasonable price. Foxwoods Casino in Ledyard, Connecticut is one example of a total solution delivered by IBM. Boy does it work. I had my first look at Foxwoods two weeks ago and was seriously impressed. World's largest casino. Not that I did any gambling. No showgirls either. Oh well.


Microsoft's approach offers ways of distributing the processing load among intelligent subsystems - if needed. Instead of building an entire tablet pc, one could just make a video card containing a VIA EDEN system and run Windows XP embedded with IE to handle almost all of the user interface display generation work. Such a card would work just fine in a pci slot, leaving the AGP bus for 3D video gaming (as an option). It would be a trivial bit of software work to pipe the output of Active Desktop to such a subsystem now that Windows has a UNIX type core. Even lower cost integrated solutions are already being worked on.


Using a future AMD Hammer system, the information could flow over a Hypertransport link to an intelligent video display system - and the intelligent general purpose video subsystem could be added to the motherboard easily. Again VIA's Centaur solutions are a leading candidate for such things. For fast 3D gaming, just pipe the drawing information to a clawhammer node that works closely with some image generation chip - essentially making an XBox core as a gaming subsystem in the main computer.


Now some skeptics might contend that HTML and XML are rather high volume descriptive languages. In a local system, HTML and XML could easily be tokenized to dramatically reduce the volume of data being moved. One table of 16-bit values probably could be used to convert ALL existing tags into 2-byte values. Suitable design work could allow using one byte values in groups where less complex descriptions were adequate. Standard desktop bit-mapped images can be stored, ready to use, in intelligent video subsystems. Or if sufficient computing power is available, vector designs can be used and scaled to any size desired. It really would be nice if the desktop icons and other graphic blandishments did scale with screen resolution instead of shrinking to nearly invisible as the resolution is set higher. Oh well.


Frankly, I have no idea whether this concept will eventually succeed or fail. Much depends on developing a certain image in the marketplace and in the minds of mass market users. As I tried to help a family that had a badly damaged install of Windows 98 - and had misplaced the Compaq system restore cds - and could barely speak English (they are from Albania), I experienced serious doubts about today's approach to marketing computer packages. The system owners were totally lost. At least they were not beating on their younger son for messing everything up. Oh well. It comes down to satisfying the users' needs first and doing it in ways that are foolproof for everyone.


So keep a lookout for more pieces of the puzzle that is being slowly assembled by Microsoft. It could work but I believe that major marketing policies and attitudes will have to change dramatically before .NET has a chance of succeeding. First, someone has to realize that gentle persuasion through truthful educational campaigns is preferable to twisting people's arms.


nils dahl

just an old man



Pssst!  We've updated our Shopping Page.