COVER FE ATURE

Human–computer systems that treat con-text simply as enumerated facts, rules, or axioms about the surrounding physical and social environment will always have trouble handling information requiring human pragmatic interpretation. One way to overcome such limitations is to inte-grate human computation in the form of microwork with tasks at which machines excel, such as big data extraction. R esearch on context awareness in computing has thus far largely focused on tasks such as identifying user location, user modes of loco-motion, and device position.1 These tasks rely on an understanding of context that is synonymous with an objective description of the physical and social envi-ronment in which devices and users are embedded.2 However, context also consists of implied knowledge