We need to move ahead of the technology curve to enable information technology research for MEMS. One way to do this would be to use today's technology (e.g. expensive sensors) to collect data that is typical of the scenarios that MEMS devices would encounter. For instance, we could use large sensors fitted with parachutes or helium-filled balloons to simulate surveillance dust. These data sets could serve as "challenge problems" for information technology research in areas such as signature analysis (Is that a truck that the smart particles just picked up, and where is it?)