I believe it should be mean absolute deviation from the mean, rather than from the next value in the list (this wasn't particularly clear in the summary or the article). So you have three numbers, mean = (1 + 2 + 10) / 3 ~= 4.3333, MAD ~= (3.333 + 2.333 + 5.666) / 3 ~= 3.778

There's another MAD, the median absolute deviation from the median, so you have for this data set median = 2, MAD = median(1, 0, 8) = 1.

Well, it's not like you're being steamrolled into accepting this as an input device. With an open platform, you should be free to use whatever input device you want.

Always introduce them to free software before the switch to Linux. It's far too much of a change to do both at the same time, and they'll reject the change entirely. Once they get used to free software on Windows, they can use the same things in the same way on Linux.

I think one of the points in saying this is what it leads to. If the upper bound is the mathematics, and the mathematics is weak (e.g. triple ROT13), then you can't get any more security than that. Well, you sort of can, but that's security through obscurity, or security theatre, which is a fairly weak stance to take.

Please explain how "download" is different from "save" from a naive computer user's point of view. In both cases, you have an object somewhere that can be considered transient / temporary, and you hope that by carrying out the operation you're transferring it to a more permanent storage medium.

Until someone can work out what the password / key is, the bitcoins will be unable to be used by anyone else -- the value of the remaining bitcoins will probably increase. If someone *is* able to work out what that password / key is, then the value of all bitcoins will drop.

Hmm, I wonder if this explains the false positive results I got when trying to find genetic markers for T1D risk (chapter 5 of my thesis).

You can make someone scared, uncertain and doubtful about your sanity by stating that they've got thousands of little hairs that they can't see on their face, and you enjoy cutting hairs that are very close to the skin using a razor blade. And then mention there might be a little blood involved, but you're game to give it a try.

My point is, it's perfectly possible to generate FUD by bending the truth (otherwise known by salespeople as "stating FACTs").

Beethoven was deaf, not blind.

I propose that Valve call this thing the Steam 'roller

I'd be interested in using this as an input device on a desktop computer. If you configure the circular pads to act like the rubber nipples, with something like a logarithmic transform to increase precision for centre movement and increase travel speed for the edges, then I can picture this being more useful than a mouse. Couple that with multi-pointer manipulation, and you get your usual pinch / zoom / rotate movements that are becoming more common as a "everyone knows how to do it" input method.

Current research (done by someone I was in biomedical science classes with) suggests that monoamine oxidase inhibitors may have a role in the increased addiction of cigarettes over plain tobacco -- although that article in particular suggests people using roll-your-own tobacco may have a harder time quitting.

The NSA, for example. Use a US server as your email service provider, and you get filtering for free!

Indiegogo allows you to set up "flexible funding" (not used for the Edge case) where all donated money goes to the project even if it doesn't reach its funding goal.

Saarland University developed a reconfigurable camera add-on, the kaleidocam which can do 3D as well as many other things. It allows you to take a single picture that is split by the device into multiple images that appear on the sensor as an array of smaller images. Possible functions include:

• Multi-spectral imaging (including simulation of different white points and source lighting)
• Light field imaging (3D, focal length change, depth of field change)
• Polarised imaging (e.g. glass stress, pictures of smoke in natural light)

Of course, this requires a single shot using a fancy lens, whereas the Harvard technique needs two frames but "no unusual hardware or fancy lenses".