Say you have thousands of people with such low-cost PCs, even shared insome netcafe, if must be. Then take the idea of wetware/human grids todo processing to analyze data. Add a payment mechanism per itemprocessed. This'd be a solution looking for a problem.
How is this for a problem:
Millions of hours of video and similar numbers of photos are takenevery day at security checkpoints such as banks, the motor vehicles/IDcard department, airports, ports, etc. This already happens widely inthe US for example. Now take a security issue like the Top Ten wantedcriminals with photos provided of each of them. How about sending allthat data soon after they are gathered to process in a wetware grid totry to identify potential suspect-matches. People are usually better atmatching faces than computers, even if it is not a perfect method. Paythe grid members a small amount per photo processed (2 to 5 centsperhaps) and ask them for sets of e.g., 100 photos to look over in asingle session; and pay a big bonus ($1000?) if an identificationeventually leads to the criminal's apprehension.
People living in many first world countries may say that $5 for aboutmaybe 1-2 hours work is simply not worth the time. But again this goesback to the "PC for the People" idea: make computing cheap for themasses in nations where $5 is a lot of money for some and $1000 is anunbelievable sum. Take a 1000 or 10,000 people in this wetware grid andyou have parallel processing of photos on a large and stillcost-effective scale.
Obviously there's a lot to be thought out here but it is the seed of an idea that could help:
- popularize the idea and utility of wetware grids
- provide a basic job for many people in developing nations
- solve complex computing problems e.g., image identification
PS: You might also want to read this month's Wired article on Crowdsourcing.