Throughout history, great teachers have created participatory environments and progressive sequences that engage learners in deep thinking, provide multiple viewpoints, support reflection, and offer frequent feedback and guidance toward higher standards. The Socratic dialogue offers one famous example in which students learn through progressive questioning from an expert. More recent social science and cognitive research has identified successful patterns in tutorial, mentoring, and group discussion interactions. Recognizing that typical chat and bulletin board systems do not organize conversation, researchers have created tools such as Knowledge Forum, CoVis, and WISE, each of which "scaffolds" learning by pre-structuring the kinds of contributions learners can make, supporting meaningful relationships among those contributions, and guiding students' browsing based on the basis of socio-cognitive and learning principles.
However, successful learning does not always fit the predefined categories supported in such tools, and does not always follow pre-identified patterns. While enabling scaffolded access to reasoning and thinking that would be too challenging for many learners without such supports, such structured access needs to be followed with more autonomous learning. Freedom to learn "outside" the structured approach is important for any given learner at different times in the course of development. Thus, we envision the field moving beyond the scaffolding of contributions and relationships, to the development of tools that allow teachers and students to dynamically and reactively structure their interactions so as to maximize future learning opportunities. For example, we need the ability to capture a session history as students work with a collaborative representation, so that the session can later be re-played. Reflection on such an animated history can encourage higher-order learning in which students become aware of their own process of learning. In addition, reviewing the history with a coach can probe critical incidents to spur deeper learning. Finally, histories might be packaged as case studies that enable others to learn more effectively. For example, a teacher might develop pedagogical insights by watching the condensed and annotated history of student interactions. How can we best capture and extract the critical moments from learning activities (video annotation, screen recording, semantic event analysis)? How can we re-use these traces to support deeper learning? How might tools support the dynamic, reactive, and reconstructive activities of expert tutors, mentors, and coaches?