In 2005, mankind created 150 exabytes of data. In 2010, we created 1,200 exabytes. The task of organizing this information is a technical challenge; the task of making meaning of this information is a social and cultural one. The Social Computing group works on models for information processing that work from both angles. We build sociotechnical tools that aim to create substantive human connections as part of the process of data analysis. Our current focus is on developing programming languages for social computation.
Research Projects
Jabberwocky
Sepandar KamvarSome of the most interesting and useful technologies of the past few years have involved large-scale coordination of people and machines. Programming languages, however, tend to focus on the machines. Traditional programming-language design assumes that people are either programmers or end-users, not members of a decentralized computing system. As a result, programming–or even thinking about–such human-machine systems is often awkward and laborious. Jabberwocky is a development stack for social computation that works to address this issue. Jabberwocky consists of Dog, a language that is intended to be executed by both computers and people, and Dormouse, a "virtual machine" layer that allows other languages and applications to be written directly to clusters of people and machines.