J. Gold, welcome to the thread.
It is also clear that Gates is well aware of the importance of embedded systems and is dedicating resource to it.
It is absolutely true that Microsoft is beginning to talk embedded systems, but it is not yet clear that Gates fully appreciates the importance of embedded systems. On a recent plane ride, I watched Gates and his chief technology officer being interviewed on TV about Microsoft's future. Both of them raved on and on about all the changes to computing just on the horizon. "If Microsoft does not evolve to meet this changes, the company will fail. The PC as we know it will not exist in 10 years", or some such statement. "The modern PC is passive, needing to be directed by people keyboarding and pointing with a mouse. In the future, computers will sense humans and respond accordingly, without humans needing to know how to instruct the computer", again loosely paraphrased. To anyone on this thread, these statements imply that Gates and therefore Microsoft see the same future we do, and our respective futures must therefore collide. Nothing could be farther from the truth.
The paraphrased statements don't imply that Microsoft's higher ups see a future of hidden ubiquitous computers, even though it sounds like they do. In Gates future, just for starts the PC will evolve to be totally capable of voice communication. This means word recognition, syntax aware, natural language capable, context sensitive, translation able and speech synthesizing. And that's just voice. The future computer will also be able to track the user's eyes and apply other physical senses to make it a better partner for humans. What kind of computer lies in Gates future? Not embedded, hidden computers. His future consists of supercomputers in PC clothing.
Gates likes the supercomputer future because his business model is based on dishing out ever more complex software at monopolistic prices. No one can compete with his complexity, or his vision, even though it will consume the lion's share of Microsoft's resources for as long as the PC paradigm survives. I find it interesting that Gates is correct in his vision, and in the direction he will be pushing Microsoft-because that's where big bucks are for foreseeable future. Gates and company simply must continually chase the $100 plus per PC licensing revenues and the $1,000 plus per server licensing revenues that will be forever available to Microsoft. At 100 million PCs annually, that's around $10 billion and growing around 16 percent annually. The 2 million Wintel servers yields another $2 billion and is growing at 30% annually. Meanwhile, as the installed base grows, software update and new application revenues add to Microsoft's fortunes.
Indeed, Microsoft's much publicized thrust into the embedded markets with Windows CE have been directed at PC-like potential markets, and not much else. (1) The handheld computers are extensions of PC's, (2) The TCI set-top box deal is an attempt to take possession of the home entertainment center console (whatever that is) and develop it into a market similar to the PC and (3) the AutoPC is an attempt to take possession of the automobile command center (whatever that is) and develop it into a market similar to the PC. Each of these markets is characterized as being in the 100 million units per year category at maturity, with unit revenues to Microsoft in the $50 to $100 or more range. Also, none of these markets requires extensive customer hand-holding, since the products are not any more heterogeneous than are today's PCs.
But what about our vision? If Gate is right, and he is, must we be wrong? And what about Windows CE invading WIND's space? Is WIND next to be eaten by Microsoft?
Although Gates has the correct vision for Microsoft, and will no doubt deliver the wonders of virtual reality, his vision is stuck in the PC paradigm, the so-called second wave of computing. The second wave will continue forth for the foreseeable future, and I for one, unlike Scott McNealy, plan to thoroughly enjoy my future superPCs. If Microsoft succeeds in imposing a Control-Alt-Del version of an AutoPC in future cars, I'll learn to enjoy them as well, rebooting at stop lights I guess. Perhaps we will all have to learn how to reboot TVs. I can live with that.
But none of this has much to do with the important things about the future of computing, the so-called third wave. Even as the Microsoft worker-bees are busy extending and complicating, and then burying complexity under complexity, the world of ubiquitous computing, communications, sensors and actuators will continue to grow exponentially at 35% to 60% annually. That world, which is diametrically opposed to Gate's world in almost every way imaginable, and growing more so, already is about the size of Gates world in units. By the middle of the next decade, the third wave numbers will be at least one, and probably two, orders of magnitude greater than Gate's world. (Of course dollars won't catch up for at least another decade after that.)
Undeniably, Windows CE 2.0 is a significant improvement over Windows CE 1.0, and version 3.0 may be better yet. However, when WIND faces Windows CE competitively, the competition isn't Microsoft, it's some no-name system integrator. The best one's are certified as Windows CE experts, and may indeed be capable. But the problem is that, unlike Microsoft's normal practice of providing shrink-wrapped software and never touching end-users, industrial embedded systems projects are heterogeneous and often focused on extracting the last dime out of unit production costs. Windows CE will be highly successful penetrating low-end, low-volume, software-centric development shops, but not the mainstream product development shops knocking out tightly engineered products at the lowest possible unit cost. The systems integrators will approach an unlimited set of embedded systems requirements with a limited set of reference designs that are often sufficient for casual applications, but rarely good enough for serious product development.
Microsoft will accept this for two reasons. The relatively high numbers of low-cost development seats will add increasingly to revenues, as will the run-time license fees. The latter will be significant because of the attractiveness of high number of projects times small run time volume times high unit run time license fee. And finally, Microsoft cannot alter its multi-billion dollar business model to address demands of small-potato, application-specific embedded systems. For example, I can't imagine Microsoft wanting to get involved in co-design or DSP operating systems. Exotic development environments are by their nature unique and not given to shrink-wrapped solutions, and therefore they conflict with Microsoft's business model. Just as Microsoft has found it difficult to succeed with business enterprise software (after years of trying, it hasn't yet dented Oracle's grip on corporate databases, and Microsoft doesn't even attempt to compete against SAP, Baan, PeopleSoft, I2 or Manugistics-all for good reasons: these are not shrink-wrapped businesses).
It seems to me that DOJ is trying to close the barn door after the horses are out. Do you think DOJ has their eye on the embedded ball?
The browser issue DOJ currently is alleging is silly, and cannot possibly impact Microsoft, its customers or end-users in any way conceivable. However, DOJ is readying additional attacks that may do damage, whether fair or not, to Microsoft, however I doubt that DOJ has any interest in the embedded systems area--Windows 98 is a more likely target. If DOJ fails to contain Microsoft under anti-trust law, Congress will change the laws to give DOJ teeth. Microsoft underestimates the obstinacy of the U.S. government because they have yet to demonstrate any knowledge whatsoever about computers. In short, irrespective of the appropriateness, effectiveness or the real of economic affects, our government is not going to let Microsoft dominate everything in software. And if the U.S. government doesn't limit Microsoft, then other governments will on a global basis. If DOJ can't solve the problem by imposing dumb rules about innocuous applications, then Microsoft will be broken apart, in as many pieces as necessary to control the monster to the satisfaction of myriad interest groups.
Should WIND "cheat" and cry foul as soon as they sense MSFT pressuring their customers and partners?
WIND would never "cheat". But WIND has a responsibility to their stakeholders (shareholders, employees, customers) to respond appropriately to behavior directed against them that they believe violates anti-trust law.
The set top box deal and the recent MSFT/APMC announcement (although the latter is small) are both of concern, but I'm sure there is more to come. Do you think that the customer set is so different from MSFT's traditional focus that these questions are moot?
The set top box deal means nothing. Not since King Solomon has the world been witness to such dismemberment. King Malone must be forcing Microsoft's or Sun's hand when he gave the server business to Sun, and the set-top box to Microsoft with the proviso that Java be used. Which Java? Used exclusively, or just available? This contract arrangement cannot work, and will probably fall apart within months. If someone gets pushed out, I would guess it would be Microsoft. The reason is the weak link in the system is on the server side, and Sun is far more experienced and capable at organizing a credible server-side solution than is Microsoft. Moreover why take the chance that Microsoft can build an adequate set-top box when they already exist from OpenTV and NCI among others.
On the other hand, if Sun gets kicked out, it still isn't a big loss for WIND. We weren't counting on it anyway. In fact, the Street has no expectations about set-top boxes for WIND. Anything that happens for WIND with set-top boxes is serendipity from the point of view of investors. (I still wonder if the 7 million-unit deal signed between NCI and Cable and Wireless involves VxWorks on each box.)
The MSFT/APMC deal is unsettling because it shows that industrial-strength tools can be fitted to Windows CE. But then again, if someone intends to use industrial-strength tools, that someone has the capability and desire to weigh all the factors when deciding on an RTOS and development toolset. Under these circumstances WIND will win far more than it looses, and that's good enough in a market becoming as sizable and pervasive as embedded systems.
Allen |
