[quoting Smith, 25 Jul 96.a] While I like the idea you are working with, I am concerned about the interaction focus you are taking. From the description, you are looking at human-computer interactions essentially in terms of interface issues: navigation, question types, etc.
I use the phrase learner-computer interaction rather than human-computer interaction to imply something a bit deeper than the interface issues. In my analysis I am looking at four aspects of the interaction:
2. Learner activities (how the learner would describe what they are doing at any given time),
3. Learner input (described in terms of the interface), and
4. System process and output (what the computer does in response to learner input or what it does automatically).
After a number of years examining Computer Assisted Learning materials from this perspective, it dawned on me that the real interaction issue wasn't with the computer, but was with the content. I would be very interested in your results if you can identify learner interactions with content and which interface mechanisms support them.
Separating the second and third aspects of the interaction, I think, allows me to look beyond the interface. However, I don't think that the interactions are content specific. For example, in Sim City the learner might be trying to patch up a road to allow traffic to pass. This type of activity is not specific to the city planning content area. A similar task could be carried out in Bihari Farmer where the learner might be trying to apply insecticide to control an outbreak of pests and thus allow the crop to grow.
Certainly you are right that describing the learners' actions in terms of their use of the interface is not sufficient. For example, in each of the above cases the learner drags an icon with the mouse, but clearly this is not sufficient to describe the interaction.