Perhaps getting the "motion kernel" is harder than I suspect it to be in a real life scenario, though.
(failure rate figure comes from http://www.anandtech.com/show/4202/the-intel-ssd-510-review/3 )
out of UC Berkely shows that doing computing with nano magnetic domains could
reduce the energy consumption by a large (million) factor."
Link to Original Source
The conventional explanation for controversy over climate change emphasizes impediments to public understanding: Limited popular knowledge of science, the inability of ordinary citizens to assess technical information, and the resulting widespread use of unreliable cognitive heuristics to assess risk. A large survey of U.S. adults (N = 1540) found little support for this account. On the whole, the most scientifically literate and numerate subjects were slightly less likely, not more, to see climate change as a serious threat than the least scientifically literate and numerate ones.
Link to Original Source
With this approach at the laboratory scale, Xu and colleagues were able to obtain a light-to-power conversion efficiency of 3.2 percent compared to 1.8 percent efficiency of conventional planar structure of the same materials.
So the efficiencies went from awful to slightly less awful.
On a related note, some of my friends thought IBM's Watson wasn't very impressive because it was receiving text input and parsing the results instead of "listening" to the answer being spoken, translating it into text, and then coming up with the question. Given my cell phone can translate speech into text, I have a feeling IBM didn't feel like that feature was important to the demonstration...
I have tried to use Arduino boards in the past, and while they're really cool for hobbyist stuff, they are very hard to integrate into battery-operated things:
1. The operating voltage is 5V (some may be 3.3V, I forget) and draw a lot of current. Batteries that supply this kind of voltage are HUGE. It would be really nice if they had a design that was optimized for low voltages and low currents, like for mobile sensing, so that I could use coin cells.
2. The devices are really memory-limited. The Uno, which is probably the most popular, has something like 2kB of ram. I used the board to interface with some sensors for tracking a flight trajectory on-board, and I could only record a few seconds of data before running out of room. Wireless transmission wasn't really an option because of power (= more batteries) limitations.
3. Connecting to USB resets the board, wiping the memory, unless you cut a trace on the board. This is supposed to help facilitate loading new programs, but becomes an annoyance if you wanted to use it to transfer sensor data stored on-board to a computer. When you cut the trace to disable the autoreset, it becomes difficult to time the reset button manually so that your program uploads.
Overall, as an EE, I was very impressed at how easy it was to use, but I think the issues I mentioned warrant some fixing if Arduino is going to be used for things like sensing.
Fast forward to me judging the high school science fair here, and I'm appalled at what the "best" these kids could muster is. Most kids couldn't even design a simple experiment. For example, one girl was measuring the conductivity of a solution and varying the temperature, but her "data" consisted of her saying that the conductivity went down as the temperature went down. There was no actual data. The best projects were judged "best" by me by at least having some kind of quantitative data, using proper controls, and having some understanding of the implications of the work. Nothing blew me away, and I had to wonder where the mentor involvement was because it seemed like these kids did everything on their own.
For background, here's the basic idea of a classical nanopore sequencer:
1. Make a solution with ions in it with a very thin membrane separating two different compartments each containing an electrode. The membrane has a very tiny hole (nanopore)
2. Apply a voltage. This will either attract or repel the salt ions, thus you get a detectable current passing through the nanopore.
3. Put DNA in the solution. The hole is hopefully small enough that the DNA can only go through as if stranded like thread through a needle. As the different base pairs move through, they block up varying amounts of the hole, manifesting as small changes in resistance across the hole.
The only real limiter is how thin you can make the membrane. Recently, some researchers used graphene, which is thinner than your average base pair, and so you do not get a resistance that is the convolution of many base pairs blocking up the pore at any given time. For more, google "Dekker DNA translocation through graphene nanopores" to see that they can already detect single pairs - and do it thousands of times a second.