You're totally on the mark, but you could take the limitation you mention and turn it into a better test. For example, instead of just asking, "what is the latency to transmit a 200MB file across 84 miles?" You could ask, "What is the relationship between message size and latency for pigeons and copper wire (or, whatever they're using) over 84 miles?" This latter test would involve transmitting files of several different sizes.

I'm betting that, at some point, the curve for pigeons will either asymptote or turn over (e.g., when the mass of 32GB microSD cards begins to exceed the flight capacity of the bird).

That's not research, that is extrapolation and interpretation (there's actually an important difference).

A touch screen is probably your best bet -- a good option would be a back-illuminated multi-touch setup so you can use thick acrylic or glass (unlikely to be broken or damaged by your toddler). If you want to go even younger, research with infants often make use to two that can be readily measures by machines: eye-gaze and suckling (on a pacifier with a sensor). Short of putting together some kind of home-brew eye-tracker, I'd suggesting hacking a pressure sensor into a pacifier. Even with that, I bet you could easily train different patterns of suckling (e.g., one pattern produces skips to the next song on the playlist, another pattern skips to the next artist).

The data might reflect something like "public interest." Any gentoo user knows that to find information related to gentoo, they should go to gentoo.org or #gentoo or the gentoo-wiki. Similar parallels can probably be drawn for debian and ubuntu. So, the data probably do not reflect the number of people using those distros but people seeking more information about them who probably not already users.

Maybe now with a billion samples, we can start training people how to recognize it.

So, basically it's a PS3 without the games and support for codecs that were published after 1992? Sweet!

Honestly, I don't get PITA's position. It's no longer part of an animal because...? Maybe because it doesn't have nerve cells that fire given "painful" stimulation? What if it does? Who's going to care about the poor little piece of meat that has to exercise all day long and experience the burn of its own lactic acid until some fat 'Merican orders it super-sized? Or maybe it's not part of an animal because there's no "brain" for the signals to reach? If that's the case, we should genetically construct brainless cows and have them running off arduinos instead. Does someone have a script for chewing I can download? But, surely someone would protest that. If only those who prefer PIC over ATMEL.

Actually, my wife was stuck in an elevator once and while her cellphone couldn't maintain a signal well enough to call out, she could text and email.

I have a Ph.D. in psychology and I can tell you that the notion of "escaping from reality" is a concept borrowed more from pop-psychology than science. The fact of the matter is, these people need exposure to cues that trigger traumatic memories in a context that is safe, supervised, and controlled. Talking about roadside bombs is an important first step but far less immersive (and less effective) for later parts of therapy than being part of an animated scene where patients get to re-experience a convoy-support mission.

I've been following a lot of similar stories recently and I don't understand why agencies and institutions wouldn't build on an opensource infrastructure that they can control (e.g., something like openlife). What happens if Linden labs goes belly-up?

Lots of research with animals and humans analyzing choice has found that changing from one option to another (e.g., the decision to move from one game zone to another) is controlled by the relative rate of reinforcement for those choices AND the effort required to change from one choice to the other. For example, studies have manipulated the physical distance between two keys in an operant chamber and found that as the distance increases, animals tend to respond more exclusively on the key with the higher-rate of reinforcement. As the cost or effort to change decreases, choice becomes less sensitive to differences in the rate of reinforcement among the alternatives. Instead of manipulating physical distance, experimenters have also required subjects to respond on a change-over key. Increases in the response-requirement on the change-over key produce effects similar to increasing physical distance. So, increasing the effort to move from one zone to another likely produces real changes in how players distribute their time in those zones. Good game designers probably distribute the best in-game rewards as an increasing function of the difficulty to obtain those rewards. It's probably also similar to having NPCs of varying difficulty. So, the question could have been, why don't MMOs just equip players with the best stuff when they first start?

Seriously though, how do they know which botnet sent which piece of spam?