I already commented it but still...Notes taken on a paper by M. Zouinar, M. Relieu, G. Calvet and P. Salembier about a new methodology to analyze collaborative mobile activities and multimodality. Constraints for designing an observation framework/method to study mobile activities:

In order to analyze mobile situations, there are two solutions: - lab situations: for instance observing a cell phone user on a running machine. The problem here is that you don't take into account elements of natural context (people who pass by for instance). - natural situations with the observer who films the pda user, the problem here is that you have a specifi interactions between the user and the observer and this is a bias.

In situations where you use multimodality (switching from one modality to another one like using text instead of audio), the context is of tremendous importance.

COnstraints are: - to preserve in an optimal way the natural character of the context in order to access to the link between people's activities and the context, turn it into something that could be analyzed. The point is to see how people change of modalities. - mobile constraints: the method should be light, and don't require too much energy (battery), not disruptive (for the user and the people who pass by).

Replay: they used logfiles enriched by users' self-confrontation (they were shown their interaction's traces like video of what they saw). This is meant to make them remenber interaction episodes and eas e 'verbal production' about their activities during the task.