There is much talk today about faith-based science. That is, should we accept as scientific that which best conforms to ancient texts or that which is observable, verifiable and repeatable? It may be tempting for the engineering community to say we are above all that. After all, we deal only with the measurable and the quantifiable. But in fact, we may be taking too much on faith. For example:
Faster is better. The original typewriter keyboard was designed for maximum efficiency. It was called the DVORAK keyboard after the configuration of letters in the top left. But the earliest typists got so proficient they'd outpace the mechanical device, and the keys jammed in the gate. So a new keyboard was created (the QWERTY system we still use) to adapt the machine to the pace of the typist. We don't do that anymore. We expect people to double their output every year because that's what machines do. Let's face it, we can longer physically keep up with our ingenuity.
Smaller is better. Yeah, kind of. But we're stuck with fingertips the size of fingertips. There's a point where miniaturization will reach a wall, and we'll have to surgically implant our digital devices if we are to keep up with Moore's Law. Tapping into our neural networks is sure to be the next step. But if having our PCs hacked seems an inconvenience, imagine having our brains hacked.
Entrepreneurship beats craftsmanship. Today's business model rewards only a very limited skill set that of the entrepreneur and those who succeed there are rewarded wonderfully. Those with other skill sets, such as craftsmen, see their jobs go offshore. How long before this bubble pops? Recall how in March 2000, investment bankers were masters of the universe; 18 months later, firemen were the new heroes. This isn't to argue that we need one or the other we need the one and the other.
Intellectual property is paramount. This goes along with the notion that entrepreneurship is superior to craftsmanship. But the people history has deemed greatest are those who freely gave away their knowledge and wisdom.
There is too a free lunch. Every technical advance grows out of a felt need, but invariably that advance creates a situation requiring a further "fix," which in turn will need yet another fix. We must get better at doing inventories of technical effects beforehand, and the first step is to acknowledge we should even undertake that effort.
"Can" is more important than "should." The idea here is that if something can be done, it should be done and if there's a market for it, it must be done. The classic case is the atomic bomb. True, it was wartime, but the decision to deploy the bomb was too far-reaching to be made by only a handful of men. For all their brilliance, Oppenheimer and his colleagues apparently never considered the folk wisdom that "what goes around comes around." One day, a weapon small enough to carry in a shopping bag could be aimed at us.
Which brings us back to faith-based science. What if those who deal in certainties decide science should not depend on theories based on relativity and uncertainty? How is it the most advanced technological society in history is even asking these questions? If quantum physics goes, where does that leave electronics? The tech community's own faith-based notions warrant reconsideration.
By Tom Mahon, a veteran technology journalist who has written two books about Silicon Valley as well as two one-man plays (firstname.lastname@example.org)
Blog Doing Math in FPGAs Tom Burke 24 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...