Taken from the interesting article you referenced:
>Processing plus access plus sensors will set the stage for the next >wave--interaction. By "interaction" we don't mean just
>Internet-variety interaction among people--we mean the
>interaction of electronic devices with the physical
>world on our behalf.
Let's look at this carefully. "Processing plus access" exactly equates to the symbiosis of ubiquitous computing and communications, forming the next major wave of computing. The "plus sensors" adds something that is buried in the notion of the "next paradigm for relating to computers". In that non-PC paradigm, the computer will interface with humans in human ways, rather than requiring humans to learn how to interface with computers as in the mainframe or PC paradigms. The human way, of course, is by using sensors that sense the analogue world of inhabited by humans.
In other words, ubiquitous, hidden computing implies, and indeed depends on, sensors of all types and variety, including those described in the article. It is probably not an exaggeration to say that sensors are the weak link in ubiquitous computing. First, they often require elaborate physics, calibrated and made robust through engineering, and then value engineered into small, inexpensive devices. It is only after sensors become small and cheap do they contribute explosively to ubiquitous computing.
Thus, most of the article is supportive totally of our vision of the next wave of computing, a vision that champions WIND. While it is true that sensors can be equipped with simple micro-controllers not requiring the equivalent of VxWorks; simple controllers give way to the need for sophisticated RTOS when multiple functions may be required simultaneously. This happens when the sensor must be monitored in regular intervals while also possibly undertaking a number of different tasks, such as communicating over a wired or wireless network.
There will be zillons of inexpensive smart sensors doing a single task expertly - not needing VxWorks. But there will be zillons of other sensors reporting to a smarter microprocessor, to do fancier things. And finally there will be zillons more sensors reporting to smarter microprocessors communicating to servers. The first will not need high-powered microprocessors; the second type may or may not (depending on complexity and relative costs of microprocessors, time-to-market, etc.); and the third type will tend almost exclusively to utilize sophisticated RTOS. That is enough to make WIND employees busy and investors happy.
Incidentally, don't forget that a parallel trend always is to supplant hardware analog/digital processing with equivalent processing on the general-purpose microprocessor - despite what Mark Brophy keeps harping about. Take an example. Voice Control Systems Inc. (VCSI) announced the other day that their voice recognition algorithms, heretofore always processed using add-in boards containing DSP chips, can now be invoked using the Pentium processor in PCs. One more hardware-only solution bites the dust, reducing the high marginal cost of providing voice recognition to zero. This means that embedded voice recognition similarly no longer needs special hardware, and can rely on a general-purpose, albeit powerful, microprocessor, perhaps opening a floodgate of possibilities for voice recognition. (Voice recognition is one example of computers using human ways to communicate.) This trend favors combining lots of functions on a general-purpose processor that might otherwise be implemented in hardware and combined more simplistically. (By the way, as soon as one algorithm inevitably moves from hardware to software, a new requirement emerges that overwhelms software, but which succumbs to a hardware solution, so I am not suggesting the end of hardware solutions.)
The article ends with a guess that future computing may evolve to work directly in analog with analog devices, obsoleting digital computers entirely. This guess is way beyond where my vision extends, and I have no comment about it, nor its implications for WIND. I'm so absorbed trying to comprehend the unlimited implications of a digital world, that to ask me to leap even to the next, analog paradigm is asking too much. When I say "WIND should do well as far as the eye can see", I suppose I should say instead, "at least until analog computing replaces digital." That is one risk factor that will not keep me awake at night.
Allen
PS - I have been searching for a pure play in sensors for quite some time, and would appreciate any suggestions for the best investment in this area. |
