sieve1234

That's really funny because something similar was done at my unversity (I think with bats, or maybe even flying foxes). I declined because although the devices were really small and they assured me that the animals did not come to harm I doubted that. It's like carrying around a big backpack.

This is getting OT, but sensor network applications and discussions often neglect the problem of device recovery and I know there are projects were sensors are not recovered and after they did their work (sensing stuff and sending it to a post-processing node) they go out of energy and become ... highly toxic waste.

I doubt that. I know there is research in sensor network (e.g. at my university) and they say they want ulta-small low-power devices but the experiments I have seen were never in need of that much small devices.

Of course there are applications were you want to go really low-power/extra-small, but in that case even a micro-controller capable of running Contiki OS is too big. What I want to say is, that the application space where you can't deploy a small machine with Linux on it BUT it is feasable to have a small machine with Contiki on it is pretty narrow.

I am unsure about this. It's big advantage is its size and that it does not require as much HW support like Linux (like a MMU). The disadvantage is that it's not a *nix and you loose the whole ecosystem (no Posix).

In my opinion the space that it occupies (the "Internet of Things"), is not well-defined and it may be probably cheeper to use something like a full-blown small computer (like the rasperry-pi) with Linux on it.

I think such sieve examples do not do languages a favour: It is slow and it's not true, that the Java code is that long. However I agree that it's nicer in Haskell, but the naive variant as given here (and often in fibonacci examples) is very slow compared to the naive version in a standard imperative language. My take: http://pastebin.com/JGnSPEaH. That is 19 vs 37 lines.