Table 3 shows a transcript from pattern-2 when the learner (L6) succeeded in reaching a goal. As shown in the transcript, the instructor (17) was looking at either the hand-gesture monitor (HM) or the in-context view (C), but which view they were watching was not clear in the videotape data, and the learner was looking at the AlgoBlock screen (AL) to check the result of his program execution. As soon as he realized that the trial was successful (when the sound effect representing the submarine's arrival to the goal (Snd) started), he said something that sounded like an expression of success and snapped his fingers in both hands. Then, he turned his face to the face-to-face view (FV). At that moment, the instructor said "Congratulations!." In this way, the boundary of a series of activities was marked interactively by those actions.
There are several relevant timings when the instructor could give congratulations: at the moment the submarine arrives at the goal, after the learner's remark, or after the sound effect. As well as these, this observation proved that the learner's action of turning his face to the instructor's face view could also provoke the instructor's reaction. In this sense, it can be said that the instructor's face view in the learner's site of pattern-2 played the role of the instructor's representative, or surrogate. This happened because in pattern-2 the face view was far enough apart from the other resources for the instructor to notice the learner's orientation to it. In this fragment, by using the face-to- face view as a surrogate, participants interactively marked the activity's boundary, and thus to achieve the end of a session successfully.
In pattern-1, by contrast, reply, reaction, and evaluation were not always immediately performed. The instructor, not knowing learner's orientation to himself/herself, often failed to catch the proper time to intervene in a learner's activity.
However, there were still some problems common to both patterns.
One of the problems was that learners could not know which views the instructor was oriented towards. There was an asymmetry in communication resources between learners and the instructor, i.e., learners could not observe the instructor's in-context view, while the instructor could do. Therefore, it sometimes happened that learners began to talk regardless of instructor's state of attention, although the trouble was restored easily by conversation. Learners tended to anticipate that the instructor was watching their actions at any time, while this was not always the case.
This asymmetry in system design arises from the difference in role between an instructor and a learner, since we presupposed an educational situation rather than equally collaborative work.
This problem may be solved by introducing a new in-context view at the learner's site just like at the instructor's site. However, adding more monitors may possibly not be as effective as we suppose, judging by the following problem.
Another problem was the instructor's confusion in utilizing communication resources; the instructor sometimes pointed outside the hand-gesture camera's capture range on the hand-gesture monitor, or even pointed at a monitor that learners could not see, and the same kind of confusions may possibly happen also at the learner's site. The reason was that it was not always clear what information was transferred to the other site, although it was possible for the instructor to know, if he/she carefully watched the in-context view. However, it could be troublesome for the instructor to have to watch another monitor in order to see how his/her own fingertip showed on the hand-gesture view in the learner's site.
A possibility is that there were too many resources to utilize, or that the relationship among various communication resources was too complicated to understand. Therefore, closer examination to validate the necessity of individual resources and/or more sophisticated integration of those resources needs to be investigated in our future work.
It is well known that in everyday communication the relative position of person and person, or person and machine, plays an important role ( Goodwin, 1981, 1995). One may assume that since communication takes place using the bodies of individuals, it depends strongly on the deployment of those bodies. If one wants collaboration through a video-mediated communication system to occur in a natural way,