In attending a meeting of the petamedia project and listening to different ideas, it is clear that a major research problem that is emerging is how to utilize enormous volumes fo video and other multimedia data produced by users. Two very interesting scenarios are related to location and events — kind of related to travel application and storytelling about an event. In the firat application, there may be lots of data contributed by users — say for Waikiki beach. Some focused on the beach, some on restaurants, some on surfers, etc. A person interested in visiting Waikiki may want to explore different aspects of the place. These user contributed videos may be used to let a user explore different aspects of Waikiki using this data. People could get a very good idea of what is available at Waikiki from non-marketing sources including about Hotels.
For the same event, say the Airport crash in New York, there may be videos and photos available from different users recording during different time intervals and from different perspectives. None of these may give a complete story of the event, but by doing a smart cit-and-paste one could tell a good story of the event.
Clearly the research problems are very challenging but signals are clear that people are going to expect tools to do this. ANd this is a good challenge for the technical community. It is nice to see that such discusions have already started in technical community and the petamedia project (disclosure: I am an external advisor to this European project).