One iteration of the exercise raised an interesting point about the personhood of the "guide" role. It was observed that the role could be supplied by not only persons but also by artefacts. A reference work could be deemed to serve as a guide. A person supplying a response to an email message for help could also serve as a guide. An automated response system could also be a guide...
Actor-Network Theory can help with the analysis of series of situations where either machines and human beings can guide participaton.
A machine can learn to recognize habits. Can a machine exercise the judgement as to when and where to give appropriate feedback based on the habits it may track? What kinds of information can machines supply human participants? For example, how much can you learn about a group of learners by comparing ther relation between the time stamps of server access to WWW materials with the time stamps of email sent to the instructor? How can such tracking features of networked environments be used to shape the system of instruction?