This paper discusses software that links data scattered across separate displays. Software intended for users who need to integrate documents from different displays, from different devices. The example given is an investigator correlating information from 1700 files--intelligence reports, news articles, charts, and pictures. The software allows the user to create visual links across displays, highlighting relationships. Data can be linked in several ways; two of these are by keywords or by a characteristic determined by the user, for example, when a person uses two different names. The software also allows for the spatial rearrangement of documents to create an insightful data presentation, for example, creating a timeline.

The paper describes how the linking lines are coded and how the client-server architecture is employed. System testing was done by eight undergraduates--each was asked to uncover terrorist plots through the analysis of 41 documents and 22 pictures. Half the testers were able to make links between the four different displays on different devices; the other half used only one display. Those using cross-device links identified somewhat more plots. They also used more displays and spent about 12 percent more time on the exercise. The debriefing comments show how users can adapt the system to serve different problem-solving strategies. As expected, being able to link many displays onto a single display made it possible for a user to integrate much information.

The paper has practical implications for people designing displays or considering person-machine interaction in general. Many stimulating ideas, but complicated sentences make it difficult to read.