Some thoughts about machine-human interface and enabling adaptability

(This article responds to Reactive Environments)

One direction of machine-human interface design
In [1], Buxton mentions Ubiquitous Media, the spirit of which I interpret as making it trivial whether people are interacting physically or virtually - whatever they need to do in a meeting they can do it. This spirit is stated in terms of the goal of using UbiComp and UbiVid in communication technology, or let's put it, human-human technology. However, the machines still exist but are made transparent or invisible. In this paper, the authors mention more than once how they focus on reducing "the cognitive burden and hence, the amount of training required", offering "an intuitive way of interacting with our surroundings", etc. This dawns on me a refreshed direction of machine-human interface design. Conventionally, we started from mainframe machine, control panel (the real one!), etc. , which point to a direction of using computer with specific and advanced operation skills and prior knowledge. Now, we want to head to the other direction, where computer can be used most naturally and intuitively as if there is no computer at all. Note that this is spoken from the users' perspective. To a research or developer, this actually means far more difficult and complicated system design and implementation.

One thought about enabling adaptability
This article raises an interesting issue of enabling system's adaptability, making it remember the characteristic of each user. This certainly belongs to the area of artificial intelligence or machine learning. Nevertheless, for video conference technology, I think it could be achieved using some simple approach. The process of a video conference can be seen as a (deterministic) finite automata (DFA). There are states like light-on/off, starting w/wo documents, speaking with high/low volume, etc. The system records one user's history of making such *decisions* and use it to guide the next adjustment to make. What's more, a system is not only able to adjust but also able to predict. For instance, if a user never uses the digital white board. The board might not even need to be switched on when the conference begins (well this also puts some risk in case one day that user suddenly wants to use it).

Reference
[1] Living in Augmented Reality: Ubiquitous Media and Reactive Environments. In K. Finn, A. Sellen & S. Wilber (Eds.). Video Mediated Communication. Hillsdale, N.J.: Erlbaum, 363-384. An earlier version of this chapter also appears in Proceedings of Imagina '95, 215-229