Assignment #8- Option #2

Visual Information as a Conversational Resource in Collaborative Physical Tasks discusses videoconferencing and compares it to audio-only and physical presence scenarios. It describes two experiments with a collaborative task, bicycle repair, and analyzes performance and language issues as it varies across the different media.
Some ideas that were expected to be relevant in each situation are the need to comply with Gricean norms and the needs to maintain situational awareness and common ground. Situational awareness is directly influenced by the amount of visual information available because it relates to maintaining a person's mental model of the environment. Common ground is affected by both language and environment because all senses can gather information to ground two people.
The experiments evaluated the effectiveness of a head-mounted video system that allowed both the worker and the helper see what the worker is looking at while repairing the bicycle. The first experiment compared solo performers and worker-helper pairs, with the latter varying between three media (audio-only and video with two types of audio). In this experiment, having video did not improve performance-either speed or accuracy-over having audio only. But the language used was different.
The second experiment combined the first one with a third state, that of physical co-presence. This side-by-side condition contained more efficient dialogue as well as better performance. It was again clear that video had shortcomings that canceled out most of its advantages.
To me, there seems to be two components to the problem of visual co-presence. One is the ability to maintain a full view of the environment that is rich enough and stable enough to closely mimic physical co-presence. The other, closely related component, is the interaction between the conversation partners, the worker and helper. If there was lots of feedback between the two and this feedback was natural and easy to provide, the visual image of the environment need not be so rich and complex.
This brings me to the point of "affordances", which Kraut et al mention on p. 21. In the HCI sense of the term, affordance is an object's natural fit or function in its environment. For example, the most natural way to hold a cup is by its handle; that is its affordance. I'm not sure how related this is to the topic of videoconferencing, but it may be worthwhile for Kraut et al to expand on their discussion of affordances. How, in a very general sense, can a videoconferencing system be made to provide very natural, common-sense tools for its users? Maybe these systems should be highly specialized by task (such as bicycle repair) or maybe collaboration can be generalized.