Certainly this stuff isn't utility fog, but it facilitates the investigation of a vast area of software development that can only now be done with simulations, giving questionable performance and questionably accurate results.

When it wakes up from the factory, a foglet must do several things. It needs to establish an identity for itself, sufficiently unique not to confuse its near neighbors. It may need to determine its physical location in the fog. It will need to participate in the self-organization of a communication network spanning the fog. There may be entirely new high-level abstractions that will be useful in programming utility fog and its kin. Gerald Sussman et al are exploring these issues.

The motes don't move, and they don't have effectors, but they start to give us an early handle on a host of complex software issues that will require non-trivial solutions. And even without movement or effectors, they are a useful (and probably necessary) step toward utility fog.

Very true. However, I personally think that although this smart dust doesn't even come close to what true utility fog can do, it's definitely a precursor. And I guess that's what I really meant. Sorry for any misunderstanding.

TanMauWu wrote on 29-may-2001:

[UC-Berkeley smart dust is] not quite true utility fog, but we're getting there.

IMHO it's very *much* not utility fog; the article doesn't mention any mechanical effectors on the "motes"; nor does the group's webpage, and given known MEMS stiction problems I'd guess Pister et al are skipping *that* challenge in favor of size reductions. Foglets need to operate in space-filling mobile lattices, but these motes are intended to remain fixed on surfaces.

(This project is only one of ~80 funded by the DARPA MEMS program.)

In concept, it seems closer to the "localizer" dust used in Vernor Vinge's novel _A Deepness in the Sky_, as mentioned in this _Wired_ article, "The Digital Gaia" and this Philcon 2000 speech.