Diversity Training
by Lindsey Vereen
In the 56K modem war that is
raging, the U.S. Robotics division of 3Com has the edge over Rockwell. U.S. Robotics is pushing its x2 technology for 56K modems, while a coalition led by Rockwell is retaliating with an incompatible K56Flex technology. Rockwell admits to being nearly 60 days behind U.S. Robotics, according to the San Francisco Chronicle, but even that brief lag gives U.S. Robotics the chance gain market share.
Time to market is just one challenge developers face. Others include increasing system complexity and a
greater dependence on software for competitive advantage. The electronics industry continues to shrink electronics products into smaller and smaller packages. Stuffing circuitry into fewer chips leads to more complex systems and tougher integration and debug challenges. Complexity is increasing even faster in software. The relative software content of electronics systems is surpassing the hardware content because of the value that software adds. To induce customers to buy their new modems, both U.S. Robotics
and Rockwell have promised free software upgrades to comply with the eventual industry standard. Increasingly, software is the solution.
One significant challenge that developers face is diversity. Unlike the PC business that has settled on a few standards, the embedded systems market has no center. In this issue, Larry Mittag says "The embedded world is difficult to characterize." In another article, Jim Turley observes that "A typical embedded system does not exist." Because embedded systems have
to meet such a broad range of needs, the kind of coalescing that occurred in the PC industry is not likely to occur in the embedded industry. It will always be more heterogeneous than the desktop market because of the diversity of roles microcontrollers play and the range of complexity you will find.
This diversity and the continued market growth will lead to new processor architectures and new tools. According to Turley, there are 100 different 32-bit chips for embedded applications right now,
and that's not the heart of the business. Moreover, the embedded tool business has years to go to before it reaches maturity. As the market increases in size, new vendors will appear, offering not just new tools, but new classes of tools designed to help you with tougher design problems that must be solved in shorter periods of time.
The increasing breadth and diversity in the embedded systems business means is that sooner or later-and probably sooner-you are likely to find yourself in the position
of evaluating architectures and tools for your next project. The idea of subjecting yourself to an evaluation process is enough to strike terror in the heart of your present vendors and perhaps you as well. Vendors don't want to lose their customers and are torn over how best to keep them. In the best of all possible worlds, competition will make better products available at attractive prices.
While changing architectures and tools is facilitated by portable languages such as C, it is still a
nontrivial effort that developers must approach with trepidation. Making such a change means cutting through hyperbole from a lot of salespeople to ferret out the best technical solution at the best price with the best chance of making you successful. The diversity of architectures and tools means fewer available standards, a greater number of choices you have to make, and the increased risk you have to take. If all goes well, you too may end up with a 60-day lead over your competitor.
