Recently the work “crowdsourcing” received a lot of attention in the tech world. The basic assumption for the success application of “crowdsourcing” is that human beings are better and faster in some tasks (e.g. image tagging) than computer. And if those tasks can be partitioned into many little tasks, which can be carried out by a single person, then “crowdsourcing” might be the right technique to solve the problem. In this sense, “crowdsourcing” is also a form of “cloud computing”, only that the “cloud” consists of people rather than computers.
There are different incentives: 1. money 2. fun 3. fame 4. team work .. and different applications can use one or multiple of those.
The criteria of problems that can be solved/benefited from “crowdsourcing” is an important topic. I have already mentioned two — 1. easy for human, hard for computer 2. parallelizable. The third criterion, is “verifiable” or “falsifiable”, i.e. there is a way or an authority verify or falsify the contribution. The way can be computation, or can be human judges.
The tooling around the crowdsourcing is also interesting. A reputation system is needed to rate the contributions, which could be anonymous. Game theory strategies might be needed to design the crowdsourcing incentive and scheme.
