...particularly physicists who think they can code.

Deal with people....

Games development on linux? Do you work for Valve?

If the US got hit by a mega tsunami it would have a lot more to worry about that a few nuclear plants getting flooded and melting down. The death toll from a mega-tsunami would probably push into the millions. Displaced people... many many more.

Anyway, even if they do melt down, the more modern reactor designs have systems for capturing and spreading the corium to reduce the possibility of uncontrolled runaway (ie it would just ruin the reactor and would not necessarily pollute the area).

Make sure you launch them into something interesting. I would recommend a raytracer - the basics are incredibly simple, but they can be expanded to great levels of complexity. There is direct visual feedback, so rather than just printing out a load of numbers the users can field like they have achieved something more substantial. Numerical optimization and data structures can be introducted gradually and immediate results can be seen. Raytracers provide a great environment for introducing object oriented programming, they are also trivial to parallelise. On top of all this raytracing is extremely useful and the knowledge gained writing a raytracer, other than the computing aspect, is extremely valuable in engineering and physics (eg the maths + physics behind them).

If you don't want to use a human standing next to the blind singer then it could be solved with some nice cheap modern technology.

If you could practically do it, attach a MEMS accelerometer (or gyro) to the baton and track the velocity of the tip (or equivalent reference point). Either wired or wirelessly transmit that velocity data to small processing box that drives a haptic device to alert the singer. Unfortunately I'm not entirely familiar with the visual clues of conducting, I've had a look at the patterns of motion on wikipedia and it seems the beat occurs at the point of zero vertical velocity following a downward stroke. If this is the case it should be relatively easy to process the velocity/acceleration information. An algorithm that has an understanding of the expected stroke would be better than a simply velocity test - it would potentially be more reliable as it will have a degree of inference, but minimizing latency may make that problematic.

In terms of haptic feedback, a sharp tap to the leg or hand would probably be better than a vibration as it has a more defined temporal position. Of course with a tap the processing algorithm has to be reliable...... alternatively a vibration could simply be engineered so that the magnitude of the vibration corresponds to the vertical position of the baton. This would mean more processing of the information by the singer, but is trivial to build electronically.

An Arduino would be perfect for the processing/driver.

He said mitigated, not prevented. I've (unintentionally) measured the oscillating light output of an incandescent while I was developing an optical trigger circuit for my last job, the intensity dropped by ~20% for this particular bulb (20W desk lamp) during the AC zero crossing. The flickering was 100Hz (funnily enough) - higher than most peoples' periphery will notice.

I think your maths suffered from one to many G&Ts..

single:

    2.00 - 0.55 = 1.45 GP (72.5% selling cost)
    1.45 - 0.50 = 0.95 NP (47.5% selling cost)

double:

    3.00 - 1.10 = 1.90 GP (63.3% selling cost)
    1.90 - 0.50 = 1.40 NP (46.7% selling cost)

erm.... yeah its not like Samsung produces billions of other chips. They must be terrified.

Good for life after transiently being really, *really* bad....

...because unlike you he obviously isn't a moron. I also didn't have to scroll last time I looked on apple.co.uk.

They could sit on them until DDR3 supporting devices get "old".