I'm sometimes asked by non-technical friends and family “what are these things they call chips [ICs]?” and “how do they work?” I give an explanation as best I can, matching it to the questioner's background and aptitude, of course. I end by saying to them that when you think about it, it's almost sheer magic that we can build these things and make them work, and work repeatedly and reliably.
But sometimes even we, as engineers, accept impressive or unexpected accomplishments without the respect they deserve. This is true even when they overcome long-held beliefs.
Two examples come to mind. Consider digital subscriber line (DSL): we were taught over and over that the local loop of the telephone line had a usable 3-dB bandwidth from about 300 to 3000 Hz. This tight bandwidth was OK because the inherent bandwidth of the loop was low; the loop had uneven, unknown, and changing spectral response; the loop was noisy; and that's the part of the voice spectrum that contains the most energy and intelligibility, so everything would be OK. Outside that relatively narrow band, there won't much to work with.
Apparently the developers of DSL didn't get the word about the narrow bandwidth, noise, and overall lousy spectral characteristics. By removing any telco-installed line filters, applying sophisticated analog and digital signal processing techniques, and adding some advanced data encoding and decoding, the local loop now supports fairly impressive data rates between the end user and the nearby central office. That's quite an accomplishment, and it didn't come easy.
For another example, look at the now-common mass-flow meter for fluids, based on the Coriolus effect (note: it is an “effect”; “force” is a misnomer). This is a very tangible physical effect which we see in large-scale, macro-level rotation of air and water currents on Earth, and the apparent curvature of the path of long-range artillery and missiles, to cite some specifics.
Somehow, however, engineers were able to take this macro-effect and harness it on a much smaller scale to build accurate, relatively compact mass flowmeters for fluids and slurries. By passing the fluid through an oscillating U-shaped tube, and sensing tube's twist due to the Coriolus effect, it's possible to make this measurement, albeit with some serious design and technical difficulties. When I first heard this type of meter described, I thought “They did what? How? Not possible!”
It looks like we should almost never say “never”, even to ourselvesand I am thinking about phasers, transporters, tricorders, cloaking devices, and even warp drive!
-x x x-