I wanted to share some of the thoughts I had on how the tagging algorithm for ‘spatialised’ conversation should work. In the following, I will offer an example of tag construction for one of the messages left into the system as the algorithm should operate. The idea is that every time that a message is posted, the algorithm should operate an analysis of the connected markers and later assign a computed tag to the message. In the next example, I will describe a situation in which Mike leaves a message in the system for Paul.
In the initial phase, Mike post the message: “Jog 2 PM?”, standing in his office. Paul retrieves the message from his office. In phase 2 Mike moves to “Hyde” park at the location where the message was posted. In the third phase, Paul joins Mike at the park. The algorithm, at this point, should “googles” a bit on WordNet, trying to find a meaning for Jog. Then it should relate this with jogging, then sport. In addition, following the indication of time that matches with the time of meeting of Paul and Mike in the park, the algorithm should also recognise that the message is about coordinating for a meeting. Finally, two tags are added to the message MSG contained in the database: “sport” and “meeting point”.