After introducing the concept of multiple monitors, which is a computer system with a physically partitioned but virtually contiguous display space (a single computer with many monitors attached), we discuss open Human-Computer Interaction multiple-monitor research areas including window management. We argue to conduct a high-level study of window management practices and a low-level study specifically comparing single-monitor and multiple-monitor window management practices. When combined with other field work on multiple monitors, the studies suggest that there is an increasingly crucial distinction between input focus (where the active window is) and user focus (where the user is actually looking on-screen) since multiple monitors encourage users to display reference information in non-active windows to aid interaction in the active window. To further explore this distinction we constructed three tools: Snip; Snap; and Mudibo. We deployed Snip and Snap in a field study, finding that participants used Snip in many of the ways that we expected though Snap did not appear to be as useful. Results from our follow-up laboratory-based study indicated that Snip can provide multiple-monitor users with dramatic time savings for referencing the snipped windows as compared to regular, overlapping windows. A laboratory-based study of Mudibo, a dialog box placement interface, provided further motivation of the tool and uncovered key interface improvements necessary to make Mudibo suitable for everyday multiple-monitor screen interaction. The findings support the original conclusion about the initial field work, namely that understanding the potentially larger gap between input focus and user focus necessitates appropriately targeted user interface development and evaluation.