A few years ago, when a speaker promised a surge in robotic applications, a skeptic
in the audience interrupted: “Robotics advocates have promised soaring applications
for fifty years and haven’t delivered. What’s different this time?” I don’t recall the
speaker’s answer, but my answer would have been: “Today, we can put millions of
transistors of intelligence in a knee joint for less than a dollar.”
Advocates for reconfigurable systems, who have a history going back to the beginnings
of Altera and Xilinx more than twenty years ago, face similar skepticism.
Times change. Robots are surging into applications—reconfigurable systems will
soon follow.
To put the conclusion up front, the engineering community is coming out of a
thirty-year stall in the development of design methods that was caused by the enormous
success of the microprocessor. The microprocessor will move from its central
role in problem solving to a supervisory role. “Paged” circuit definition will displace
instruction-based algorithms as the workhorse of systems. For the engineering community,
it will be a reluctant transition, one that will be forced by the rapidly growing
market for consumer mobile devices.