Untangling the usability of fisheye menus

Fisheye menus have become a prominent example of fisheye interfaces, yet contain several nonfisheye elements and have not been systematically evaluated. This study investigates whether fisheye menus are useful, and tries to untangle the impact on usability of the following properties of fisheye menus: use of distortion, index of letters for coarse navigation, and the focus-lock mode for accurate movement. Twelve participants took part in an experiment comparing fisheye menus with three alternative menu designs across known-item and browsing tasks, as well as across alphabetical and categorical menu structures. The results show that for finding known items, conventional hierarchical menus are the most accurate and by far the fastest. In addition, participants rate the hierarchical menu as more satisfying than fisheye and multifocus menus, but do not consistently prefer any one menu. For browsing tasks, the menus neither differ with respect to accuracy nor selection time. Eye-movement data show that participants make little use of nonfocus regions of the fisheye menu, though these are a defining feature of fisheye interfaces. Nonfocus regions are used more with the multifocus menu, which enlarges important menu items in these regions. With the hierarchical menu, participants make shorter fixations and have shorter scanpaths, suggesting lower requirements for mental activity and visual search. We conclude by discussing why fisheye menus are inferior to the hierarchical menu and how both may be improved.