Notes taken at the meeting with my phd advisor: Main topic of my phd: study of collaboration and collaborative processes (CSCW, in mobile settings) -> We want to investigate whether and how spatial information (location?) impacts ONE collaborative processes which is Mutual Modeling (also called intersubjectivity).
The main assumption is that space is meaningfull in terms of mutual modeling: - from the POSITION of his/her partner, one can infer his/her ACTIVITY - from the DIRECTION of his/her partner, one can infer his/her INTENTIONS/STRATEGY
Keywords: mobile computing, awareness, mutual modeling, collaboration, CSCW
The thesis would include three components (that should be perceived as the main scientific contributions): - a semi-controlled experiment - development of innovative methods to analyze spatial interactions, performance as well as group interaction and mutual modelling. - model or ontology of the agents actions + interactions (inter-agents as well as agents/world).
About the model, 2 possibilities: - ontology, classification, structured description, modeling coordination or MM... - predictive modelling (much more difficult)
I see this as EXPERIMENT -> MODEL OF AGENTS (so as to describe the agents + their interactions or to improve the external validity of the model drawn from the experiments...)
Concerning the task we want to adress, there must be a mapping between the REAL SPACE and the PROBLEM SPACE. It could be solving a murder party (cluedo-like where people have to investigate in various rooms to solve the problem) or some diagnostic-like task.
In terms of experimental stuff, the global hypothesis is that SPATIAL INFORMATION impacts MUTUAL MODELING and GROUP MODELLING (improve it). "Spatial information" means here spatial awareness and mainly location + direction awareness.
Hence we can test: - with location awareness tool/without - with synchronous awareness tool/asynchronous (discriminate presence/location/direction) - ...
Notes: - showing that people use schelling point to solve the task ? - from this, we can fairly imagine lots of "laboratory tasks" like showing movies of a person's trajectory on an EPFL map and asking subjects to say what he is doing, where he is heading... some inferences. - the REPLAY dimension is essential ! It would be very interesting to show the participants (of CatchBob! for instance) their trajectory on the map and then ask them to comment it, to describe episodes