Kamen originally approached Quadir after learning of his programs to finance individual, usually poor, entrepreneurs. Through Quadir's existing microloan program, called GrameenPhone, 175,000 Bangladeshi women have taken out loans of about $200 to buy cell phones and subscriptions, which they then sell time on to local villagers. Since 1976, Grameen Bank has loaned out $5 billion to the poor, in amounts as small as $25 for investments in livestock, or as large as several hundred dollars for housing space, which can also be used for small-scale manufacturing. The bank has recovered 99 percent of its loans, according to its website.

But the generator project can't go anywhere until Kamen and Quadir find a manufacturer to produce the device. "It's very hard to get big companies to start putting serious money toward tooling a product when they are not sure that the ultimate purchaser has the resources and wherewithal to pay for the product," Kamen says.

The cell phone program, says Quadir, was successful in part because cheap handsets were already being produced in large quantities. Kamen hopes that the encouraging results of trial in Bangladesh may help convince companies to invest.

Large-scale programs designed to bring water to millions at a time depend on building extensive infrastructure. The advantage of Kamen's generator and water purifier, he argues, is their flexibility. They can be carried into remote villages by just two people. And the generator can run on any local fuel available.

The design concept behind the water purifier, like the Stirling engine, is nothing new: it works by heating and distilling water. What does make it unique is its efficiency -- the generator reclaims about 98% of the heat normally lost in the distillation process and reuses it to distill more water, Kamen says.

Running continuously on a few hundred watts of power, a single purifier should provide enough water for a village of 100 people, Kamen estimates. And it can purify water from any source, regardless of what contaminants it might contain. That rules out the need for quality testing or specialized treatment.

Amy Smith points out, however, that to use the water purifiers effectively, villagers will also need some education. Often, inadequate storage methods mean water is contaminated after it's gathered, Smith says.

If the system works, and people know how to use it, Quadir believes the potential benefits can go beyond better health. A distributed water system owned by local entrepreneurs will employ more people than a centralized system, as Quadir's experience with the cell phone program proved to him.

He likewise expects that distributed power, and water purification, will mean more jobs -- and perhaps more democracy. Says Quadir, "Technology allows people to rise from below."