In the conception paper, I described interaction structures as a communication medium with role specifications and instructions about how to interact with each other.
A communication medium serves as tool that learners can convey meanings into speeches or texts when interacting with each other. During the daily talk that occurs face-to-face or in speeches, we are seldom aware that we actually rely spoken language as a tool to communicate except when we encounter trouble to express ourselves. When communicating in a computer-mediated environment such as chats and wikis, it is more difficult to express and clarify what we mean when other non-verbal clues are gone. Therefore, supporting learners how to effectively use communication media is an important part of designing the interaction structures unless we are certain that the users are already capable of using them without difficulties.
As for role tasks, from the current experiment I observed that not all learners were aware of their role tasks or were able to perform them even everyone was assigned to act as tutor or tutee with instructions of role tasks. One reason might be that these participants were inexperienced in tutoring or peer learning approaches. If this is true, then it is necessary to train or guide learners prior to the interaction process.
Regarding structuring interaction, numerous studies in peer tutoring and collaborative learning have emphasized that interactions must be scaffold, guided or even trained prior to implementation. Yet, in Learning Networks the degree of these structuring interventions should be at a different level and suit different purposes. The ideal goal of social learning is to build knowledge through interactions. However, knowledge sharing or even merely information exchange might occur mostly in Learning Networks. How to design interaction structures that suit for all levels of knowledge acquisition/learning is the current challenge for me.